FINAL -- LSTM approach with added weekends.ipynb


# -*- coding: utf-8 -*-
# In[1]:


import numpy as np , pandas as pd
from data_pipeline import transformation_pipeline
from sklearn.model_selection import train_test_split
import tensorflow as tf


# In[2]:


data = pd.read_csv('content/preprocessed_train.csv')


# In[3]:


pipeline, data_cleaned= transformation_pipeline(data, building_id=122 , meter=0 , primary_use= 99 )


# In[4]:


transformed_data= pipeline.fit_transform(data_cleaned)
display( pd.DataFrame(transformed_data , index=data_cleaned.index , columns= data_cleaned.columns ).head() )


# In[12]:


display( transformed_data[:,0] ) # this gives us the meter reading


# In[13]:


display(transformed_data[:, 1:]) # this gives us the rest of the columns


# In[8]:


x_train,x_val,y_train,y_val= train_test_split(  transformed_data[:, 1:] ,
                                                transformed_data[:,0] ,
                                                test_size=0.2,
                                                shuffle=False,
                                                random_state=2021)


# In[9]:


train_gen= tf.keras.preprocessing.sequence.TimeseriesGenerator( x_train,
                                                                y_train ,
                                                                length=6 ,sampling_rate=1,
                                                                stride=1, batch_size=32
                                                                )

val_gen= tf.keras.preprocessing.sequence.TimeseriesGenerator(   x_val ,
                                                                y_val  ,
                                                                length=6 ,sampling_rate=1,
                                                                stride=1, batch_size=32
                                                                )


# In[30]:


model= tf.keras.Sequential([tf.keras.layers.LSTM(128 , activation='relu' ,return_sequences=False),
        tf.keras.layers.Dense(1) ] )


# In[31]:


model.compile(loss='mse', optimizer= tf.keras.optimizers.SGD(0.0001))

#cb= tf.keras.callbacks.ModelCheckpoint('models/LSTM_per_hour_model', monitor='val_loss', verbose=0, save_best_only=True)
# Fitting the model
history=model.fit(train_gen,
                  validation_data= val_gen,
                  epochs=300 ,
                  #callbacks=[cb],
                  shuffle=False)


# In[40]:


display(val_gen[7][1]) # where 7 is the batch , 0 stands for the features and 1 stands for the output


# In[93]:


predicted_batch_7 = model.predict (val_gen[7][0] )


# In[16]:


import matplotlib.pyplot as plt
_, ax=plt.subplots(figsize=(10,5))
ax.plot(range(32),
         predicted_batch_7,
        color='green',label='Predicted')

ax.plot( range(32),
         val_gen[7][1],
        color='red',label='Actual')
ax.legend()

plt.show()
# In[32]:


# lets try predicting more than one patch

predicted=[]
actual=[]
for i in range(32):
    predicted.extend(model.predict (val_gen[i][0] )  )
    actual.extend( val_gen[i][1])


# In[33]:


fig , (ax1 , ax2, ax) = plt.subplots( 3,1 ,  figsize=(30,15), sharex=True )

ax1.plot(range(len(actual)) ,
         predicted,
         color='green')
plt.legend()

ax2.plot( range(len(actual)) ,
          actual,
          color='red')
plt.legend()

ax.plot(range(len(actual)) ,
         predicted,
         color='green',
           label='Predicted')
plt.legend()
ax.plot( range(len(actual)) ,
          actual,
          color='red',
           label='actual')

plt.legend()

plt.title('Test_set', loc='center')

plt.show()
# In[98]:


## Let's try to see the effect on the training data


# In[20]:


predicted_t=[]
actual_t=[]
for i in range(32):
    predicted_t.extend(model.predict (train_gen[i][0] )  )
    actual_t.extend( train_gen[i][1])


# In[21]:


fig , (ax1 , ax2, ax) = plt.subplots( 3,1 ,  figsize=(30,15), sharex=True )

ax1.plot(range(len(actual_t)) ,
         predicted_t,
         color='green')

ax2.plot( range(len(actual_t)) ,
          actual_t,
          color='red')


ax.plot(range(len(actual_t)) ,
         predicted_t,
         color='green',
           label='Predicted')

ax.plot( range(len(actual_t)) ,
          actual_t,
          color='red',
           label='actual')
plt.title('Traine_set', loc='center')

plt.legend()

plt.show()
# In[ ]: