notebooks/covid/evaluate_benchmark_medium.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "killing-nitrogen",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "\n",
    "# Set paths\n",
    "sys.path.extend([\n",
    "    os.path.abspath(os.path.join(\"../..\")),\n",
    "    os.path.abspath(os.path.join(\"../../../BayesFlow_dev/BayesFlow/\"))\n",
    "])\n",
    "\n",
    "\n",
    "import time\n",
    "from functools import partial\n",
    "import pickle\n",
    "import bayesflow as bf\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from simulation import setup_simulator\n",
    "from network import setup_network\n",
    "from helpers import get_metrics, calculate_setting_times"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7c96a630",
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "57d161b4",
   "metadata": {},
   "source": [
    "# Constants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "3d9d6759",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "16384\n"
     ]
    }
   ],
   "source": [
    "RNG = np.random.default_rng(2023)\n",
    "\n",
    "SIM_BUDGET = 2**14\n",
    "print(SIM_BUDGET)\n",
    "NUM_EPOCHS = 75\n",
    "NUM_ENSEMBLE_MEMBERS = 2\n",
    "BUDGET_SETTING = f\"{SIM_BUDGET}_budget\"\n",
    "ALPHA_VALUES = [0.5, 1.0, 2.0]\n",
    "config = {\"T\": 14, \"N\": 83e6}\n",
    "CHECKPOINT_FOLDER = f\"checkpoints/benchmark/{BUDGET_SETTING}\"\n",
    "TIME_SETTINGS = [1, 3, 1000]\n",
    "TRAINING_TIMES_PATH =f\"{CHECKPOINT_FOLDER}/training_times.pkl\"\n",
    "COLOR_CODES = ['#35b779', '#31688e', '#440154'] # one for each alpha setting"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8cbc7241",
   "metadata": {},
   "source": [
    "# Evaluate networks"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92c8fa45",
   "metadata": {},
   "source": [
    "## Get posterior draws"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "24542616",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Loaded loss history from checkpoints/benchmark/16384_budget/powerscaled/net0\\history_75.pkl.\n",
      "INFO:root:Networks loaded from checkpoints/benchmark/16384_budget/powerscaled/net0\\ckpt-75\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference times for powerscaled_net0: [1.4703388214111328, 1.471337080001831, 1.4703385829925537]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Loaded loss history from checkpoints/benchmark/16384_budget/powerscaled/net1\\history_75.pkl.\n",
      "INFO:root:Networks loaded from checkpoints/benchmark/16384_budget/powerscaled/net1\\ckpt-75\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference times for powerscaled_net1: [1.4919962882995605, 1.3131945133209229, 1.3051881790161133]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Loaded loss history from checkpoints/benchmark/16384_budget/unscaled/alpha_0.5/net0\\history_75.pkl.\n",
      "INFO:root:Networks loaded from checkpoints/benchmark/16384_budget/unscaled/alpha_0.5/net0\\ckpt-75\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference times for unscaled_alpha_0.5_net0: [1.3432214260101318, 1.416689395904541, 1.2821669578552246]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Loaded loss history from checkpoints/benchmark/16384_budget/unscaled/alpha_0.5/net1\\history_75.pkl.\n",
      "INFO:root:Networks loaded from checkpoints/benchmark/16384_budget/unscaled/alpha_0.5/net1\\ckpt-75\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference times for unscaled_alpha_0.5_net1: [1.4112827777862549, 1.3121931552886963, 1.181074619293213]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Loaded loss history from checkpoints/benchmark/16384_budget/unscaled/alpha_1.0/net0\\history_75.pkl.\n",
      "INFO:root:Networks loaded from checkpoints/benchmark/16384_budget/unscaled/alpha_1.0/net0\\ckpt-75\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference times for unscaled_alpha_1.0_net0: [1.1760711669921875, 1.0379457473754883, 1.1680629253387451]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Loaded loss history from checkpoints/benchmark/16384_budget/unscaled/alpha_1.0/net1\\history_75.pkl.\n",
      "INFO:root:Networks loaded from checkpoints/benchmark/16384_budget/unscaled/alpha_1.0/net1\\ckpt-75\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference times for unscaled_alpha_1.0_net1: [1.0409464836120605, 1.1450417041778564, 1.253141164779663]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Loaded loss history from checkpoints/benchmark/16384_budget/unscaled/alpha_2.0/net0\\history_75.pkl.\n",
      "INFO:root:Networks loaded from checkpoints/benchmark/16384_budget/unscaled/alpha_2.0/net0\\ckpt-75\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference times for unscaled_alpha_2.0_net0: [1.2601428031921387, 1.3742499351501465, 1.2961785793304443]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Loaded loss history from checkpoints/benchmark/16384_budget/unscaled/alpha_2.0/net1\\history_75.pkl.\n",
      "INFO:root:Networks loaded from checkpoints/benchmark/16384_budget/unscaled/alpha_2.0/net1\\ckpt-75\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference times for unscaled_alpha_2.0_net1: [1.2971796989440918, 1.2861695289611816, 1.368243932723999]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'powerscaled_net0': 1.4706714948018391,\n",
       " 'powerscaled_net1': 1.3701263268788655,\n",
       " 'unscaled_alpha_0.5_net0': 1.347359259923299,\n",
       " 'unscaled_alpha_0.5_net1': 1.301516850789388,\n",
       " 'unscaled_alpha_1.0_net0': 1.127359946568807,\n",
       " 'unscaled_alpha_1.0_net1': 1.1463764508565266,\n",
       " 'unscaled_alpha_2.0_net0': 1.3101904392242432,\n",
       " 'unscaled_alpha_2.0_net1': 1.3171977202097576}"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get draws & inference times from each network for all three alpha settings\n",
    "draws_dict = {}\n",
    "inference_times = {}\n",
    "\n",
    "def load_data(factor):\n",
    "    return pickle.load(open(f'simulations/sims_{factor}.pkl', 'rb+'))\n",
    "\n",
    "for i in range(len(ALPHA_VALUES) + 1): # Loop over 3 unscaled networks + already powerscaled network\n",
    "    \n",
    "    if i == 0: # Powerscaled network\n",
    "        network_name = \"powerscaled\"\n",
    "        checkpoint_base_path = f\"{CHECKPOINT_FOLDER}/powerscaled\"\n",
    "        is_fix_alpha = False\n",
    "        prior, simulator, model = setup_simulator(config=config)\n",
    "    else: # Unscaled networks\n",
    "        alpha_net = ALPHA_VALUES[i-1]\n",
    "        network_name = f\"unscaled_alpha_{alpha_net}\"\n",
    "        checkpoint_base_path = f\"{CHECKPOINT_FOLDER}/unscaled/alpha_{alpha_net}\"\n",
    "        is_fix_alpha = True\n",
    "        prior, simulator, model = setup_simulator(config=config, fix_alpha=ALPHA_VALUES[i-1])\n",
    "\n",
    "    draws_dict[network_name] = {}\n",
    "\n",
    "    for m in range(NUM_ENSEMBLE_MEMBERS):\n",
    "        network_num = f\"net{m}\"\n",
    "        checkpoint_path = f\"{checkpoint_base_path}/{network_num}\"\n",
    "\n",
    "        #prior_means, prior_stds = load_standardization_params(STANDARDIZATION_PATH, network_name) # use alpha of net, not setting for configuration\n",
    "        amortizer, trainer = setup_network(\n",
    "            generative_model=model, \n",
    "            checkpoint_path=checkpoint_path,\n",
    "            is_fix_alpha=is_fix_alpha\n",
    "        )\n",
    "        \n",
    "        draws_dict[network_name][network_num] = {}\n",
    "        inference_times_per_setting = []\n",
    "\n",
    "        for alpha_setting in ALPHA_VALUES: # Loop over all alpha settings\n",
    "            alpha_str = str(alpha_setting)[0] + str(alpha_setting)[-1] # use some funny stuff to reuse ALPHA_VALUES (remove dot to get from 0.5 to 05 etc)\n",
    "            conf_data = trainer.configurator(load_data(alpha_str))\n",
    "            _ = amortizer.sample(\n",
    "                conf_data, \n",
    "                n_samples=250\n",
    "            ) # \"warmup\" - first draw is slower, so times would be biased by network position in loop otherwise \n",
    "            start_time = time.time()\n",
    "            draws_dict[network_name][network_num][f\"conf_{alpha_str}\"] = amortizer.sample(\n",
    "                conf_data, \n",
    "                n_samples=2500 # use many samples to get stable estimates, especially of posterior contraction\n",
    "            )\n",
    "            end_time = time.time() \n",
    "            inference_times_per_setting.append(end_time - start_time)\n",
    "\n",
    "        print(f\"Inference times for {network_name}_{network_num}: {inference_times_per_setting}\")\n",
    "        \n",
    "        # Get average time over alpha settings per network\n",
    "        inference_times[f\"{network_name}_{network_num}\"] = np.mean(inference_times_per_setting) \n",
    "inference_times"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7216e7a2",
   "metadata": {},
   "source": [
    "## Calculate metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "c81424a1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Data Scaling</th>\n",
       "      <th>Network Scaling</th>\n",
       "      <th>RMSE</th>\n",
       "      <th>MAE</th>\n",
       "      <th>R2</th>\n",
       "      <th>Correlation</th>\n",
       "      <th>ECE</th>\n",
       "      <th>Posterior Contraction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.5</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>23.303855</td>\n",
       "      <td>8.705355</td>\n",
       "      <td>0.410984</td>\n",
       "      <td>0.63143</td>\n",
       "      <td>0.009421</td>\n",
       "      <td>0.497702</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.5</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>22.792439</td>\n",
       "      <td>8.615815</td>\n",
       "      <td>0.405494</td>\n",
       "      <td>0.616779</td>\n",
       "      <td>0.019211</td>\n",
       "      <td>0.482752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>23.881284</td>\n",
       "      <td>8.920368</td>\n",
       "      <td>0.425119</td>\n",
       "      <td>0.636187</td>\n",
       "      <td>0.005026</td>\n",
       "      <td>0.420081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>24.023817</td>\n",
       "      <td>8.897533</td>\n",
       "      <td>0.429548</td>\n",
       "      <td>0.63876</td>\n",
       "      <td>0.007447</td>\n",
       "      <td>0.406111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.5</td>\n",
       "      <td>1.0</td>\n",
       "      <td>21.380071</td>\n",
       "      <td>7.901129</td>\n",
       "      <td>0.327064</td>\n",
       "      <td>0.620657</td>\n",
       "      <td>0.105789</td>\n",
       "      <td>0.707749</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.5</td>\n",
       "      <td>1.0</td>\n",
       "      <td>24.169924</td>\n",
       "      <td>7.927007</td>\n",
       "      <td>0.334368</td>\n",
       "      <td>0.614558</td>\n",
       "      <td>0.106263</td>\n",
       "      <td>0.707861</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.5</td>\n",
       "      <td>2.0</td>\n",
       "      <td>22.015698</td>\n",
       "      <td>8.300046</td>\n",
       "      <td>0.175283</td>\n",
       "      <td>0.569873</td>\n",
       "      <td>0.199895</td>\n",
       "      <td>0.862863</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.5</td>\n",
       "      <td>2.0</td>\n",
       "      <td>22.013945</td>\n",
       "      <td>8.326355</td>\n",
       "      <td>0.168618</td>\n",
       "      <td>0.57191</td>\n",
       "      <td>0.200895</td>\n",
       "      <td>0.869751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>15.21077</td>\n",
       "      <td>5.564077</td>\n",
       "      <td>0.414647</td>\n",
       "      <td>0.591644</td>\n",
       "      <td>0.007658</td>\n",
       "      <td>0.357131</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>14.89631</td>\n",
       "      <td>5.498874</td>\n",
       "      <td>0.414429</td>\n",
       "      <td>0.59139</td>\n",
       "      <td>0.004947</td>\n",
       "      <td>0.349752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>20.595906</td>\n",
       "      <td>7.171581</td>\n",
       "      <td>0.361707</td>\n",
       "      <td>0.592445</td>\n",
       "      <td>0.040263</td>\n",
       "      <td>-0.180131</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>20.657516</td>\n",
       "      <td>7.177172</td>\n",
       "      <td>0.365736</td>\n",
       "      <td>0.594467</td>\n",
       "      <td>0.053737</td>\n",
       "      <td>-0.181717</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>14.607633</td>\n",
       "      <td>5.388227</td>\n",
       "      <td>0.40844</td>\n",
       "      <td>0.587584</td>\n",
       "      <td>0.008289</td>\n",
       "      <td>0.380339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>15.540338</td>\n",
       "      <td>5.451065</td>\n",
       "      <td>0.412297</td>\n",
       "      <td>0.58932</td>\n",
       "      <td>0.007368</td>\n",
       "      <td>0.372263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>13.343948</td>\n",
       "      <td>4.897174</td>\n",
       "      <td>0.33143</td>\n",
       "      <td>0.569625</td>\n",
       "      <td>0.106763</td>\n",
       "      <td>0.689763</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>13.327279</td>\n",
       "      <td>4.90496</td>\n",
       "      <td>0.327053</td>\n",
       "      <td>0.569328</td>\n",
       "      <td>0.103263</td>\n",
       "      <td>0.697864</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>2.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>9.5705</td>\n",
       "      <td>3.672866</td>\n",
       "      <td>0.386951</td>\n",
       "      <td>0.512691</td>\n",
       "      <td>0.013158</td>\n",
       "      <td>0.140245</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>2.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>9.086169</td>\n",
       "      <td>3.526044</td>\n",
       "      <td>0.386818</td>\n",
       "      <td>0.51617</td>\n",
       "      <td>0.005842</td>\n",
       "      <td>0.279413</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>2.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>20.594229</td>\n",
       "      <td>6.849437</td>\n",
       "      <td>0.031949</td>\n",
       "      <td>0.484592</td>\n",
       "      <td>0.054737</td>\n",
       "      <td>-1.89611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>2.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>21.435455</td>\n",
       "      <td>7.015689</td>\n",
       "      <td>0.036232</td>\n",
       "      <td>0.486768</td>\n",
       "      <td>0.060737</td>\n",
       "      <td>-2.002236</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>12.750532</td>\n",
       "      <td>4.647335</td>\n",
       "      <td>0.34915</td>\n",
       "      <td>0.506211</td>\n",
       "      <td>0.047289</td>\n",
       "      <td>-0.546248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>15.773339</td>\n",
       "      <td>4.747119</td>\n",
       "      <td>0.348325</td>\n",
       "      <td>0.509213</td>\n",
       "      <td>0.051211</td>\n",
       "      <td>-0.548411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.353053</td>\n",
       "      <td>3.605674</td>\n",
       "      <td>0.383948</td>\n",
       "      <td>0.516396</td>\n",
       "      <td>0.0045</td>\n",
       "      <td>0.20326</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.394996</td>\n",
       "      <td>3.61921</td>\n",
       "      <td>0.384594</td>\n",
       "      <td>0.51259</td>\n",
       "      <td>0.005763</td>\n",
       "      <td>0.224977</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Data Scaling Network Scaling       RMSE       MAE        R2 Correlation   \n",
       "0           0.5     powerscaled  23.303855  8.705355  0.410984     0.63143  \\\n",
       "1           0.5     powerscaled  22.792439  8.615815  0.405494    0.616779   \n",
       "2           0.5             0.5  23.881284  8.920368  0.425119    0.636187   \n",
       "3           0.5             0.5  24.023817  8.897533  0.429548     0.63876   \n",
       "4           0.5             1.0  21.380071  7.901129  0.327064    0.620657   \n",
       "5           0.5             1.0  24.169924  7.927007  0.334368    0.614558   \n",
       "6           0.5             2.0  22.015698  8.300046  0.175283    0.569873   \n",
       "7           0.5             2.0  22.013945  8.326355  0.168618     0.57191   \n",
       "8           1.0     powerscaled   15.21077  5.564077  0.414647    0.591644   \n",
       "9           1.0     powerscaled   14.89631  5.498874  0.414429     0.59139   \n",
       "10          1.0             0.5  20.595906  7.171581  0.361707    0.592445   \n",
       "11          1.0             0.5  20.657516  7.177172  0.365736    0.594467   \n",
       "12          1.0             1.0  14.607633  5.388227   0.40844    0.587584   \n",
       "13          1.0             1.0  15.540338  5.451065  0.412297     0.58932   \n",
       "14          1.0             2.0  13.343948  4.897174   0.33143    0.569625   \n",
       "15          1.0             2.0  13.327279   4.90496  0.327053    0.569328   \n",
       "16          2.0     powerscaled     9.5705  3.672866  0.386951    0.512691   \n",
       "17          2.0     powerscaled   9.086169  3.526044  0.386818     0.51617   \n",
       "18          2.0             0.5  20.594229  6.849437  0.031949    0.484592   \n",
       "19          2.0             0.5  21.435455  7.015689  0.036232    0.486768   \n",
       "20          2.0             1.0  12.750532  4.647335   0.34915    0.506211   \n",
       "21          2.0             1.0  15.773339  4.747119  0.348325    0.509213   \n",
       "22          2.0             2.0   9.353053  3.605674  0.383948    0.516396   \n",
       "23          2.0             2.0   9.394996   3.61921  0.384594     0.51259   \n",
       "\n",
       "         ECE Posterior Contraction  \n",
       "0   0.009421              0.497702  \n",
       "1   0.019211              0.482752  \n",
       "2   0.005026              0.420081  \n",
       "3   0.007447              0.406111  \n",
       "4   0.105789              0.707749  \n",
       "5   0.106263              0.707861  \n",
       "6   0.199895              0.862863  \n",
       "7   0.200895              0.869751  \n",
       "8   0.007658              0.357131  \n",
       "9   0.004947              0.349752  \n",
       "10  0.040263             -0.180131  \n",
       "11  0.053737             -0.181717  \n",
       "12  0.008289              0.380339  \n",
       "13  0.007368              0.372263  \n",
       "14  0.106763              0.689763  \n",
       "15  0.103263              0.697864  \n",
       "16  0.013158              0.140245  \n",
       "17  0.005842              0.279413  \n",
       "18  0.054737              -1.89611  \n",
       "19  0.060737             -2.002236  \n",
       "20  0.047289             -0.546248  \n",
       "21  0.051211             -0.548411  \n",
       "22    0.0045               0.20326  \n",
       "23  0.005763              0.224977  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Compute metrics\n",
    "conf_names = [\"conf_05\", \"conf_10\", \"conf_20\"]\n",
    "net_list = [\"powerscaled\", \"unscaled_alpha_0.5\", \"unscaled_alpha_1.0\", \"unscaled_alpha_2.0\"]\n",
    "# Get better & consistent names\n",
    "conf_mapping = {\"conf_05\": \"0.5\", \"conf_10\": \"1.0\", \"conf_20\": \"2.0\"} \n",
    "net_mapping = {\"powerscaled\": \"powerscaled\", \"unscaled_alpha_0.5\": \"0.5\", \"unscaled_alpha_1.0\": \"1.0\", \"unscaled_alpha_2.0\": \"2.0\"}\n",
    "\n",
    "metrics_names = ['RMSE', 'MAE', 'R2', 'Correlation', 'ECE', 'Posterior Contraction']\n",
    "\n",
    "num_configs = len(conf_names)\n",
    "num_networks = len(net_list) * NUM_ENSEMBLE_MEMBERS\n",
    "num_metrics = len(metrics_names)\n",
    "\n",
    "data = np.empty((num_configs * num_networks, num_metrics + 2), dtype=object)\n",
    "# Beware: dtype object leads to some weird errors when doing further computations based on the table, use .astype(float) to fix\n",
    "\n",
    "for i, conf in enumerate(conf_names):\n",
    "    parameters = load_data(conf[-2:])['prior_draws'] # Get ground-truth parameters of setting\n",
    "    for j, net in enumerate(net_list):\n",
    "        for m in range(NUM_ENSEMBLE_MEMBERS):\n",
    "            network_num = f\"net{m}\"\n",
    "            draws = draws_dict[net][network_num][conf]\n",
    "\n",
    "            rmse, mae, r2, corr, ece, cont = get_metrics(parameters, draws)\n",
    "            \n",
    "            new_conf_name = conf_mapping[conf]\n",
    "            new_net_name = net_mapping[net]\n",
    "            \n",
    "            row_index = i * num_networks + j * NUM_ENSEMBLE_MEMBERS + m\n",
    "            data[row_index] = [new_conf_name, new_net_name, rmse, mae, r2, corr, ece, cont]\n",
    "columns = ['Data Scaling', 'Network Scaling'] + metrics_names\n",
    "benchmark_results = pd.DataFrame(data, columns=columns)\n",
    "benchmark_results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2626ce1b",
   "metadata": {},
   "source": [
    "## Visualize/aggregate results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "206845f8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x1000 with 18 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualize results\n",
    "fig, axes = plt.subplots(num_configs, num_metrics, figsize=(16, 10))\n",
    "data_scalings = list(conf_mapping.values())\n",
    "\n",
    "for i, ds in enumerate(data_scalings):\n",
    "    for j, metric in enumerate(metrics_names):\n",
    "        ax = axes[i, j]\n",
    "        subset_df = benchmark_results[benchmark_results['Data Scaling'] == ds]\n",
    "\n",
    "        colors = ['grey'] * len(subset_df)\n",
    "        # Highlight powerscaled network + unscaled network trained for the respective data scaling\n",
    "        for index, index_name in enumerate(subset_df.loc[:, 'Network Scaling']):\n",
    "            if index_name == 'powerscaled':\n",
    "                colors[index] = \"#800000\"\n",
    "            if index_name == ds:\n",
    "                colors[index] = \"#000080\"\n",
    "\n",
    "        subset_df.plot(kind='bar', x='Network Scaling', y=metric, ax=ax, alpha=0.7, color=colors, legend=False)\n",
    "\n",
    "        ax.set_xticklabels(ax.get_xticklabels(), rotation=45, ha=\"right\")\n",
    "        ax.set_axisbelow(True)\n",
    "        ax.grid(True, alpha=0.3)\n",
    "\n",
    "    # Add data scaling name to the very left of each row\n",
    "    axes[i, 0].set_ylabel(f'Data Scaling {ds}', rotation=90, size=20)\n",
    "\n",
    "# Add metric names at the very top of each column\n",
    "for j, metric in enumerate(metrics_names):\n",
    "    axes[0, j].set_title(metric, size=20)\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "4358c7b9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">RMSE</th>\n",
       "      <th colspan=\"2\" halign=\"left\">MAE</th>\n",
       "      <th colspan=\"2\" halign=\"left\">R2</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Correlation</th>\n",
       "      <th colspan=\"2\" halign=\"left\">ECE</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Posterior Contraction</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Net category</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>powerscaled</th>\n",
       "      <td>15.053540</td>\n",
       "      <td>0.222357</td>\n",
       "      <td>5.531475</td>\n",
       "      <td>0.046105</td>\n",
       "      <td>0.414538</td>\n",
       "      <td>0.000154</td>\n",
       "      <td>0.591517</td>\n",
       "      <td>0.000179</td>\n",
       "      <td>0.006303</td>\n",
       "      <td>0.001917</td>\n",
       "      <td>0.353442</td>\n",
       "      <td>0.005218</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unscaled</th>\n",
       "      <td>15.073985</td>\n",
       "      <td>0.659522</td>\n",
       "      <td>5.419646</td>\n",
       "      <td>0.044433</td>\n",
       "      <td>0.410368</td>\n",
       "      <td>0.002727</td>\n",
       "      <td>0.588452</td>\n",
       "      <td>0.001227</td>\n",
       "      <td>0.007829</td>\n",
       "      <td>0.000651</td>\n",
       "      <td>0.376301</td>\n",
       "      <td>0.005711</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   RMSE                 MAE                  R2             \n",
       "                   mean       std      mean       std      mean       std   \n",
       "Net category                                                                \n",
       "powerscaled   15.053540  0.222357  5.531475  0.046105  0.414538  0.000154  \\\n",
       "unscaled      15.073985  0.659522  5.419646  0.044433  0.410368  0.002727   \n",
       "\n",
       "             Correlation                 ECE           Posterior Contraction   \n",
       "                    mean       std      mean       std                  mean   \n",
       "Net category                                                                   \n",
       "powerscaled     0.591517  0.000179  0.006303  0.001917              0.353442  \\\n",
       "unscaled        0.588452  0.001227  0.007829  0.000651              0.376301   \n",
       "\n",
       "                        \n",
       "                   std  \n",
       "Net category            \n",
       "powerscaled   0.005218  \n",
       "unscaled      0.005711  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Aggregate results for benchmark table in paper\n",
    "\n",
    "# Select only powerscaled or matching unscaled networks\n",
    "benchmark_results_matching = benchmark_results[\n",
    "    (benchmark_results['Network Scaling'] == 'powerscaled') | \n",
    "    ((benchmark_results['Data Scaling'] == benchmark_results['Network Scaling']) & \n",
    "    (benchmark_results['Network Scaling'] != 'powerscaled'))\n",
    "].copy()\n",
    "\n",
    "# Categorize as powerscaled or unscaled\n",
    "benchmark_results_matching['Net category'] = (\n",
    "    benchmark_results_matching['Network Scaling']\n",
    "    .apply(lambda x: 'powerscaled' if x == 'powerscaled' else 'unscaled')\n",
    ")\n",
    "\n",
    "agg_dict = {metric: ['mean', 'std'] for metric in metrics_names}\n",
    "benchmark_results_matching_baseline = benchmark_results_matching[benchmark_results_matching['Data Scaling'] == '1.0']\n",
    "benchmark_results_agg = benchmark_results_matching_baseline.groupby('Net category').agg(agg_dict).astype(float)\n",
    "benchmark_results_agg"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "96b93a53",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Data Scaling</th>\n",
       "      <th>Network Scaling</th>\n",
       "      <th>RMSE</th>\n",
       "      <th>MAE</th>\n",
       "      <th>R2</th>\n",
       "      <th>Correlation</th>\n",
       "      <th>ECE</th>\n",
       "      <th>Posterior Contraction</th>\n",
       "      <th>Net category</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>15.21077</td>\n",
       "      <td>5.564077</td>\n",
       "      <td>0.414647</td>\n",
       "      <td>0.591644</td>\n",
       "      <td>0.007658</td>\n",
       "      <td>0.357131</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>14.89631</td>\n",
       "      <td>5.498874</td>\n",
       "      <td>0.414429</td>\n",
       "      <td>0.59139</td>\n",
       "      <td>0.004947</td>\n",
       "      <td>0.349752</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>14.607633</td>\n",
       "      <td>5.388227</td>\n",
       "      <td>0.40844</td>\n",
       "      <td>0.587584</td>\n",
       "      <td>0.008289</td>\n",
       "      <td>0.380339</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>15.540338</td>\n",
       "      <td>5.451065</td>\n",
       "      <td>0.412297</td>\n",
       "      <td>0.58932</td>\n",
       "      <td>0.007368</td>\n",
       "      <td>0.372263</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Data Scaling Network Scaling       RMSE       MAE        R2 Correlation   \n",
       "8           1.0     powerscaled   15.21077  5.564077  0.414647    0.591644  \\\n",
       "9           1.0     powerscaled   14.89631  5.498874  0.414429     0.59139   \n",
       "12          1.0             1.0  14.607633  5.388227   0.40844    0.587584   \n",
       "13          1.0             1.0  15.540338  5.451065  0.412297     0.58932   \n",
       "\n",
       "         ECE Posterior Contraction Net category  \n",
       "8   0.007658              0.357131  powerscaled  \n",
       "9   0.004947              0.349752  powerscaled  \n",
       "12  0.008289              0.380339     unscaled  \n",
       "13  0.007368              0.372263     unscaled  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "benchmark_results_matching_baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "abb49c84",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Data Scaling</th>\n",
       "      <th>Network Scaling</th>\n",
       "      <th>RMSE</th>\n",
       "      <th>MAE</th>\n",
       "      <th>R2</th>\n",
       "      <th>Correlation</th>\n",
       "      <th>ECE</th>\n",
       "      <th>Posterior Contraction</th>\n",
       "      <th>Net category</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.5</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>23.303855</td>\n",
       "      <td>8.705355</td>\n",
       "      <td>0.410984</td>\n",
       "      <td>0.63143</td>\n",
       "      <td>0.009421</td>\n",
       "      <td>0.497702</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.5</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>22.792439</td>\n",
       "      <td>8.615815</td>\n",
       "      <td>0.405494</td>\n",
       "      <td>0.616779</td>\n",
       "      <td>0.019211</td>\n",
       "      <td>0.482752</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>23.881284</td>\n",
       "      <td>8.920368</td>\n",
       "      <td>0.425119</td>\n",
       "      <td>0.636187</td>\n",
       "      <td>0.005026</td>\n",
       "      <td>0.420081</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>24.023817</td>\n",
       "      <td>8.897533</td>\n",
       "      <td>0.429548</td>\n",
       "      <td>0.63876</td>\n",
       "      <td>0.007447</td>\n",
       "      <td>0.406111</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>15.21077</td>\n",
       "      <td>5.564077</td>\n",
       "      <td>0.414647</td>\n",
       "      <td>0.591644</td>\n",
       "      <td>0.007658</td>\n",
       "      <td>0.357131</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>14.89631</td>\n",
       "      <td>5.498874</td>\n",
       "      <td>0.414429</td>\n",
       "      <td>0.59139</td>\n",
       "      <td>0.004947</td>\n",
       "      <td>0.349752</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>14.607633</td>\n",
       "      <td>5.388227</td>\n",
       "      <td>0.40844</td>\n",
       "      <td>0.587584</td>\n",
       "      <td>0.008289</td>\n",
       "      <td>0.380339</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>15.540338</td>\n",
       "      <td>5.451065</td>\n",
       "      <td>0.412297</td>\n",
       "      <td>0.58932</td>\n",
       "      <td>0.007368</td>\n",
       "      <td>0.372263</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>2.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>9.5705</td>\n",
       "      <td>3.672866</td>\n",
       "      <td>0.386951</td>\n",
       "      <td>0.512691</td>\n",
       "      <td>0.013158</td>\n",
       "      <td>0.140245</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>2.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>9.086169</td>\n",
       "      <td>3.526044</td>\n",
       "      <td>0.386818</td>\n",
       "      <td>0.51617</td>\n",
       "      <td>0.005842</td>\n",
       "      <td>0.279413</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.353053</td>\n",
       "      <td>3.605674</td>\n",
       "      <td>0.383948</td>\n",
       "      <td>0.516396</td>\n",
       "      <td>0.0045</td>\n",
       "      <td>0.20326</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.394996</td>\n",
       "      <td>3.61921</td>\n",
       "      <td>0.384594</td>\n",
       "      <td>0.51259</td>\n",
       "      <td>0.005763</td>\n",
       "      <td>0.224977</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Data Scaling Network Scaling       RMSE       MAE        R2 Correlation   \n",
       "0           0.5     powerscaled  23.303855  8.705355  0.410984     0.63143  \\\n",
       "1           0.5     powerscaled  22.792439  8.615815  0.405494    0.616779   \n",
       "2           0.5             0.5  23.881284  8.920368  0.425119    0.636187   \n",
       "3           0.5             0.5  24.023817  8.897533  0.429548     0.63876   \n",
       "8           1.0     powerscaled   15.21077  5.564077  0.414647    0.591644   \n",
       "9           1.0     powerscaled   14.89631  5.498874  0.414429     0.59139   \n",
       "12          1.0             1.0  14.607633  5.388227   0.40844    0.587584   \n",
       "13          1.0             1.0  15.540338  5.451065  0.412297     0.58932   \n",
       "16          2.0     powerscaled     9.5705  3.672866  0.386951    0.512691   \n",
       "17          2.0     powerscaled   9.086169  3.526044  0.386818     0.51617   \n",
       "22          2.0             2.0   9.353053  3.605674  0.383948    0.516396   \n",
       "23          2.0             2.0   9.394996   3.61921  0.384594     0.51259   \n",
       "\n",
       "         ECE Posterior Contraction Net category  \n",
       "0   0.009421              0.497702  powerscaled  \n",
       "1   0.019211              0.482752  powerscaled  \n",
       "2   0.005026              0.420081     unscaled  \n",
       "3   0.007447              0.406111     unscaled  \n",
       "8   0.007658              0.357131  powerscaled  \n",
       "9   0.004947              0.349752  powerscaled  \n",
       "12  0.008289              0.380339     unscaled  \n",
       "13  0.007368              0.372263     unscaled  \n",
       "16  0.013158              0.140245  powerscaled  \n",
       "17  0.005842              0.279413  powerscaled  \n",
       "22    0.0045               0.20326     unscaled  \n",
       "23  0.005763              0.224977     unscaled  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "benchmark_results_matching"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "1270a18a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">RMSE</th>\n",
       "      <th colspan=\"2\" halign=\"left\">MAE</th>\n",
       "      <th colspan=\"2\" halign=\"left\">R2</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Correlation</th>\n",
       "      <th colspan=\"2\" halign=\"left\">ECE</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Posterior Contraction</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Net category</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>powerscaled</th>\n",
       "      <td>15.05</td>\n",
       "      <td>0.22</td>\n",
       "      <td>5.53</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.41</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.59</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.35</td>\n",
       "      <td>0.01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unscaled</th>\n",
       "      <td>15.07</td>\n",
       "      <td>0.66</td>\n",
       "      <td>5.42</td>\n",
       "      <td>0.04</td>\n",
       "      <td>0.41</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.59</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.38</td>\n",
       "      <td>0.01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               RMSE         MAE          R2      Correlation        ECE        \n",
       "               mean   std  mean   std  mean  std        mean  std  mean  std   \n",
       "Net category                                                                   \n",
       "powerscaled   15.05  0.22  5.53  0.05  0.41  0.0        0.59  0.0  0.01  0.0  \\\n",
       "unscaled      15.07  0.66  5.42  0.04  0.41  0.0        0.59  0.0  0.01  0.0   \n",
       "\n",
       "             Posterior Contraction        \n",
       "                              mean   std  \n",
       "Net category                              \n",
       "powerscaled                   0.35  0.01  \n",
       "unscaled                      0.38  0.01  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "benchmark_results_agg.round(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "f1033fbe",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">RMSE</th>\n",
       "      <th colspan=\"2\" halign=\"left\">MAE</th>\n",
       "      <th colspan=\"2\" halign=\"left\">R2</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Correlation</th>\n",
       "      <th colspan=\"2\" halign=\"left\">ECE</th>\n",
       "      <th colspan=\"2\" halign=\"left\">Posterior Contraction</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Net category</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>powerscaled</th>\n",
       "      <td>15.054</td>\n",
       "      <td>0.222</td>\n",
       "      <td>5.531</td>\n",
       "      <td>0.046</td>\n",
       "      <td>0.415</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.592</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.006</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.353</td>\n",
       "      <td>0.005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unscaled</th>\n",
       "      <td>15.074</td>\n",
       "      <td>0.660</td>\n",
       "      <td>5.420</td>\n",
       "      <td>0.044</td>\n",
       "      <td>0.410</td>\n",
       "      <td>0.003</td>\n",
       "      <td>0.588</td>\n",
       "      <td>0.001</td>\n",
       "      <td>0.008</td>\n",
       "      <td>0.001</td>\n",
       "      <td>0.376</td>\n",
       "      <td>0.006</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                RMSE           MAE            R2        Correlation          \n",
       "                mean    std   mean    std   mean    std        mean    std   \n",
       "Net category                                                                 \n",
       "powerscaled   15.054  0.222  5.531  0.046  0.415  0.000       0.592  0.000  \\\n",
       "unscaled      15.074  0.660  5.420  0.044  0.410  0.003       0.588  0.001   \n",
       "\n",
       "                ECE        Posterior Contraction         \n",
       "               mean    std                  mean    std  \n",
       "Net category                                             \n",
       "powerscaled   0.006  0.002                 0.353  0.005  \n",
       "unscaled      0.008  0.001                 0.376  0.006  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "benchmark_results_agg.round(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "d9ca574f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Data Scaling</th>\n",
       "      <th>Network Scaling</th>\n",
       "      <th>RMSE</th>\n",
       "      <th>MAE</th>\n",
       "      <th>R2</th>\n",
       "      <th>Correlation</th>\n",
       "      <th>ECE</th>\n",
       "      <th>Posterior Contraction</th>\n",
       "      <th>Net category</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.5</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>23.303855</td>\n",
       "      <td>8.705355</td>\n",
       "      <td>0.410984</td>\n",
       "      <td>0.63143</td>\n",
       "      <td>0.009421</td>\n",
       "      <td>0.497702</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.5</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>22.792439</td>\n",
       "      <td>8.615815</td>\n",
       "      <td>0.405494</td>\n",
       "      <td>0.616779</td>\n",
       "      <td>0.019211</td>\n",
       "      <td>0.482752</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>23.881284</td>\n",
       "      <td>8.920368</td>\n",
       "      <td>0.425119</td>\n",
       "      <td>0.636187</td>\n",
       "      <td>0.005026</td>\n",
       "      <td>0.420081</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>24.023817</td>\n",
       "      <td>8.897533</td>\n",
       "      <td>0.429548</td>\n",
       "      <td>0.63876</td>\n",
       "      <td>0.007447</td>\n",
       "      <td>0.406111</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>15.21077</td>\n",
       "      <td>5.564077</td>\n",
       "      <td>0.414647</td>\n",
       "      <td>0.591644</td>\n",
       "      <td>0.007658</td>\n",
       "      <td>0.357131</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>14.89631</td>\n",
       "      <td>5.498874</td>\n",
       "      <td>0.414429</td>\n",
       "      <td>0.59139</td>\n",
       "      <td>0.004947</td>\n",
       "      <td>0.349752</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>14.607633</td>\n",
       "      <td>5.388227</td>\n",
       "      <td>0.40844</td>\n",
       "      <td>0.587584</td>\n",
       "      <td>0.008289</td>\n",
       "      <td>0.380339</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>15.540338</td>\n",
       "      <td>5.451065</td>\n",
       "      <td>0.412297</td>\n",
       "      <td>0.58932</td>\n",
       "      <td>0.007368</td>\n",
       "      <td>0.372263</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>2.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>9.5705</td>\n",
       "      <td>3.672866</td>\n",
       "      <td>0.386951</td>\n",
       "      <td>0.512691</td>\n",
       "      <td>0.013158</td>\n",
       "      <td>0.140245</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>2.0</td>\n",
       "      <td>powerscaled</td>\n",
       "      <td>9.086169</td>\n",
       "      <td>3.526044</td>\n",
       "      <td>0.386818</td>\n",
       "      <td>0.51617</td>\n",
       "      <td>0.005842</td>\n",
       "      <td>0.279413</td>\n",
       "      <td>powerscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.353053</td>\n",
       "      <td>3.605674</td>\n",
       "      <td>0.383948</td>\n",
       "      <td>0.516396</td>\n",
       "      <td>0.0045</td>\n",
       "      <td>0.20326</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.394996</td>\n",
       "      <td>3.61921</td>\n",
       "      <td>0.384594</td>\n",
       "      <td>0.51259</td>\n",
       "      <td>0.005763</td>\n",
       "      <td>0.224977</td>\n",
       "      <td>unscaled</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Data Scaling Network Scaling       RMSE       MAE        R2 Correlation   \n",
       "0           0.5     powerscaled  23.303855  8.705355  0.410984     0.63143  \\\n",
       "1           0.5     powerscaled  22.792439  8.615815  0.405494    0.616779   \n",
       "2           0.5             0.5  23.881284  8.920368  0.425119    0.636187   \n",
       "3           0.5             0.5  24.023817  8.897533  0.429548     0.63876   \n",
       "8           1.0     powerscaled   15.21077  5.564077  0.414647    0.591644   \n",
       "9           1.0     powerscaled   14.89631  5.498874  0.414429     0.59139   \n",
       "12          1.0             1.0  14.607633  5.388227   0.40844    0.587584   \n",
       "13          1.0             1.0  15.540338  5.451065  0.412297     0.58932   \n",
       "16          2.0     powerscaled     9.5705  3.672866  0.386951    0.512691   \n",
       "17          2.0     powerscaled   9.086169  3.526044  0.386818     0.51617   \n",
       "22          2.0             2.0   9.353053  3.605674  0.383948    0.516396   \n",
       "23          2.0             2.0   9.394996   3.61921  0.384594     0.51259   \n",
       "\n",
       "         ECE Posterior Contraction Net category  \n",
       "0   0.009421              0.497702  powerscaled  \n",
       "1   0.019211              0.482752  powerscaled  \n",
       "2   0.005026              0.420081     unscaled  \n",
       "3   0.007447              0.406111     unscaled  \n",
       "8   0.007658              0.357131  powerscaled  \n",
       "9   0.004947              0.349752  powerscaled  \n",
       "12  0.008289              0.380339     unscaled  \n",
       "13  0.007368              0.372263     unscaled  \n",
       "16  0.013158              0.140245  powerscaled  \n",
       "17  0.005842              0.279413  powerscaled  \n",
       "22    0.0045               0.20326     unscaled  \n",
       "23  0.005763              0.224977     unscaled  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "benchmark_results_matching"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "35d4150d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"3\" halign=\"left\">Time in minutes</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>3</th>\n",
       "      <th>1000</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>powerscaled</th>\n",
       "      <td>5.500668</td>\n",
       "      <td>5.54235</td>\n",
       "      <td>26.320864</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unscaled</th>\n",
       "      <td>5.557153</td>\n",
       "      <td>16.67146</td>\n",
       "      <td>5557.153378</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Time in minutes                       \n",
       "                       1         3            1000\n",
       "powerscaled        5.500668   5.54235    26.320864\n",
       "unscaled           5.557153  16.67146  5557.153378"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get training times\n",
    "\n",
    "with open(TRAINING_TIMES_PATH, \"rb\") as f:\n",
    "    training_times = pickle.load(f)\n",
    "\n",
    "# Filter out training times for unscaled networks in the 0.5 and 2.0 setting\n",
    "training_times = {key: value for key, value in training_times.items() if '0.5' not in key and '2.0' not in key}\n",
    "\n",
    "time_data = {\n",
    "    'powerscaled': calculate_setting_times(TIME_SETTINGS, training_times, inference_times, 'powerscaled'),\n",
    "    'unscaled': calculate_setting_times(TIME_SETTINGS, training_times, inference_times, 'unscaled')\n",
    "}\n",
    "\n",
    "time_results = pd.DataFrame(time_data).T\n",
    "time_results.columns = pd.MultiIndex.from_product([['Time in minutes'], TIME_SETTINGS])\n",
    "time_results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "938b234b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'powerscaled_net0': 329.377783536911,\n",
       " 'powerscaled_net1': 328.20140171051025,\n",
       " 'unscaled_alpha_1.0_net0': 329.02800464630127,\n",
       " 'unscaled_alpha_1.0_net1': 335.6049211025238}"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "training_times"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "f7699eaf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"3\" halign=\"left\">Time in minutes</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>3</th>\n",
       "      <th>1000</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>powerscaled</th>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>26.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unscaled</th>\n",
       "      <td>6.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>5557.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Time in minutes              \n",
       "                       1     3       1000\n",
       "powerscaled             6.0   6.0    26.0\n",
       "unscaled                6.0  17.0  5557.0"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "time_results.round(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f7b30038",
   "metadata": {},
   "source": [
    "## Additional exploration of results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "51cf216c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>RMSE</th>\n",
       "      <th>MAE</th>\n",
       "      <th>R2</th>\n",
       "      <th>Correlation</th>\n",
       "      <th>ECE</th>\n",
       "      <th>Posterior Contraction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>RMSE</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.921522</td>\n",
       "      <td>-0.253484</td>\n",
       "      <td>0.559772</td>\n",
       "      <td>0.346058</td>\n",
       "      <td>-0.005416</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MAE</th>\n",
       "      <td>0.921522</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.202425</td>\n",
       "      <td>0.666698</td>\n",
       "      <td>0.336897</td>\n",
       "      <td>0.114207</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>R2</th>\n",
       "      <td>-0.253484</td>\n",
       "      <td>-0.202425</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.501111</td>\n",
       "      <td>-0.583830</td>\n",
       "      <td>0.592155</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Correlation</th>\n",
       "      <td>0.559772</td>\n",
       "      <td>0.666698</td>\n",
       "      <td>0.501111</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.013953</td>\n",
       "      <td>0.650048</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ECE</th>\n",
       "      <td>0.346058</td>\n",
       "      <td>0.336897</td>\n",
       "      <td>-0.583830</td>\n",
       "      <td>0.013953</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.220843</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Posterior Contraction</th>\n",
       "      <td>-0.005416</td>\n",
       "      <td>0.114207</td>\n",
       "      <td>0.592155</td>\n",
       "      <td>0.650048</td>\n",
       "      <td>0.220843</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           RMSE       MAE        R2  Correlation       ECE   \n",
       "RMSE                   1.000000  0.921522 -0.253484     0.559772  0.346058  \\\n",
       "MAE                    0.921522  1.000000 -0.202425     0.666698  0.336897   \n",
       "R2                    -0.253484 -0.202425  1.000000     0.501111 -0.583830   \n",
       "Correlation            0.559772  0.666698  0.501111     1.000000  0.013953   \n",
       "ECE                    0.346058  0.336897 -0.583830     0.013953  1.000000   \n",
       "Posterior Contraction -0.005416  0.114207  0.592155     0.650048  0.220843   \n",
       "\n",
       "                       Posterior Contraction  \n",
       "RMSE                               -0.005416  \n",
       "MAE                                 0.114207  \n",
       "R2                                  0.592155  \n",
       "Correlation                         0.650048  \n",
       "ECE                                 0.220843  \n",
       "Posterior Contraction               1.000000  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Correlation matrix: Agreement between metrics\n",
    "metric_corr_matrix = np.corrcoef(benchmark_results.loc[:, metrics_names].astype(float).transpose())\n",
    "pd.DataFrame(metric_corr_matrix, columns=metrics_names, index=metrics_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "47f1a6f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Recovery for all combinations\n",
    "for i, conf in enumerate(conf_names):\n",
    "    parameters = load_data(conf[-2:])['prior_draws']\n",
    "    for net in net_list:\n",
    "        for m in range(NUM_ENSEMBLE_MEMBERS):\n",
    "            network_num = f\"net{m}\"\n",
    "            f = bf.diagnostics.plot_recovery(\n",
    "                draws_dict[net][network_num][conf_names[i]],\n",
    "                parameters,\n",
    "                param_names=[\"Recovery\", \"for\", f\"data scaling {data_scalings[i]}\", \"and\", f\"{net} net\"], # Funny title hack\n",
    "                color=COLOR_CODES[i]\n",
    "            )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d7eb559b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Calibration for all combinations\n",
    "for i, conf in enumerate(conf_names):\n",
    "    parameters = load_data(conf[-2:])['prior_draws']\n",
    "    for net in net_list:\n",
    "        for m in range(NUM_ENSEMBLE_MEMBERS):\n",
    "            network_num = f\"net{m}\"\n",
    "            f = bf.diagnostics.plot_sbc_ecdf(\n",
    "                draws_dict[net][network_num][conf_names[i]],\n",
    "                parameters,\n",
    "                param_names=[\"Calibration\", \"for\", f\"data scaling {data_scalings[i]}\", \"and\", f\"{net} net\"], # Funny title hack\n",
    "                rank_ecdf_color=COLOR_CODES[i],\n",
    "                difference = True\n",
    "            )"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.10"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": true,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}