Integrate review changes

NC0DER · Aug 9, 2020 · 0020e22 · 0020e22
1 parent 4d20639
commit 0020e22
Show file tree

Hide file tree

Showing 21 changed files with 196,984 additions and 66 deletions.
diff --git a/CORD19_GraphOfDocs/datasets/dataset1/fully_balanced/test_balanced_840.csv b/CORD19_GraphOfDocs/datasets/dataset1/fully_balanced/test_balanced_840.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset1/fully_balanced/train_balanced_668.csv b/CORD19_GraphOfDocs/datasets/dataset1/fully_balanced/train_balanced_668.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset2/fully_balanced/test_balanced_1566.csv b/CORD19_GraphOfDocs/datasets/dataset2/fully_balanced/test_balanced_1566.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset2/fully_balanced/train_balanced_858.csv b/CORD19_GraphOfDocs/datasets/dataset2/fully_balanced/train_balanced_858.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset3/fully_balanced/test_balanced_2636.csv b/CORD19_GraphOfDocs/datasets/dataset3/fully_balanced/test_balanced_2636.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset3/fully_balanced/train_balanced_1726.csv b/CORD19_GraphOfDocs/datasets/dataset3/fully_balanced/train_balanced_1726.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset4/fully_balanced/test_balanced_7798.csv b/CORD19_GraphOfDocs/datasets/dataset4/fully_balanced/test_balanced_7798.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset4/fully_balanced/train_balanced_3346.csv b/CORD19_GraphOfDocs/datasets/dataset4/fully_balanced/train_balanced_3346.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset5/fully_balanced/test_balanced_12976.csv b/CORD19_GraphOfDocs/datasets/dataset5/fully_balanced/test_balanced_12976.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset5/fully_balanced/train_balanced_5042.csv b/CORD19_GraphOfDocs/datasets/dataset5/fully_balanced/train_balanced_5042.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset6/fully_balanced/test_balanced_16276.csv b/CORD19_GraphOfDocs/datasets/dataset6/fully_balanced/test_balanced_16276.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset6/fully_balanced/train_balanced_5296.csv b/CORD19_GraphOfDocs/datasets/dataset6/fully_balanced/train_balanced_5296.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset7/fully_balanced/test_balanced_25900.csv b/CORD19_GraphOfDocs/datasets/dataset7/fully_balanced/test_balanced_25900.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset7/fully_balanced/train_balanced_6210.csv b/CORD19_GraphOfDocs/datasets/dataset7/fully_balanced/train_balanced_6210.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset8/fully_balanced/test_balanced_34586.csv b/CORD19_GraphOfDocs/datasets/dataset8/fully_balanced/test_balanced_34586.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset8/fully_balanced/train_balanced_8578.csv b/CORD19_GraphOfDocs/datasets/dataset8/fully_balanced/train_balanced_8578.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset9/fully_balanced/test_balanced_49236.csv b/CORD19_GraphOfDocs/datasets/dataset9/fully_balanced/test_balanced_49236.csv
diff --git a/CORD19_GraphOfDocs/datasets/dataset9/fully_balanced/train_balanced_13034.csv b/CORD19_GraphOfDocs/datasets/dataset9/fully_balanced/train_balanced_13034.csv
diff --git a/cd_plot_alrotf.py b/cd_plot_alrotf.py
@@ -0,0 +1,69 @@
+#!pip install orange3
+import Orange
+import matplotlib.pyplot as plt
+from google.colab import files
+import pandas as pd
+
+def run_test():
+    names = [
+        'logit_comb_1',
+        'logit_comb_2',
+        'logit_comb_3',
+        'logit_comb_4',
+        'knn_50_comb_1',
+        'knn_50_comb_2',
+        'knn_50_comb_3',
+        'lsvm_comb_1',
+        'lsvm_comb_2',
+        'lsvm_comb_3',
+        'lsvm_comb_4',
+        'svm_comb_1',
+        'svm_comb_2',
+        'svm_comb_3',
+        'svm_comb_4',
+        'nn_comb_1',
+        'nn_comb_2',
+        'nn_comb_3',
+        'nn_comb_4',
+        'nn_comb_5',
+    ]
+
+    avranks = [
+        17.22222222222222,
+        14.333333333333332,
+        3.5000000000000004,
+        5.222222222222221,
+        16.77777777777778,
+        13.222222222222223,
+        5.888888888888888,
+        14.722222222222221,
+        11.000000000000002,
+        1.888888888888889,
+        5.0,
+        15.88888888888889,
+        13.666666666666668,
+        7.888888888888888,
+        12.833333333333334,
+        15.555555555555557,
+        14.277777777777777,
+        6.9444444444444455,
+        12.166666666666668,
+        12.166666666666668
+    ]
+
+    pairs = zip(names, avranks)
+    pairs = sorted(pairs, key=lambda x: x[1])
+    for i in pairs:
+        print(i)
+
+
+
+    eos = len(names)
+    print(len(names), len(avranks))
+    cd = Orange.evaluation.compute_CD(avranks[0:eos], 90 , alpha = '0.05' , test = 'nemenyi' )#'bonferroni-dunn')# test='nemenyi') 
+    print(cd)
+    Orange.evaluation.graph_ranks(avranks[0:eos], names[0:eos], cd = cd, width = 12, textspace = 2.5)
+    a = plt.show()
+    #plt.savefig('labeled_ration_5%.png', format='png', dpi=600) 
+
+run_test()
diff --git a/experiments.ipynb b/experiments.ipynb
diff --git a/ranks_and_cd.txt b/ranks_and_cd.txt
@@ -0,0 +1,23 @@
+('lsvm_comb_3', 1.888888888888889)
+('logit_comb_3', 3.5000000000000004)
+('lsvm_comb_4', 5.0)
+('logit_comb_4', 5.222222222222221)
+('knn_50_comb_3', 5.888888888888888)
+('nn_comb_3', 6.9444444444444455)
+('svm_comb_3', 7.888888888888888)
+('lsvm_comb_2', 11.000000000000002)
+('nn_comb_4', 12.166666666666668)
+('nn_comb_5', 12.166666666666668)
+('svm_comb_4', 12.833333333333334)
+('knn_50_comb_2', 13.222222222222223)
+('svm_comb_2', 13.666666666666668)
+('nn_comb_2', 14.277777777777777)
+('logit_comb_2', 14.333333333333332)
+('lsvm_comb_1', 14.722222222222221)
+('nn_comb_1', 15.555555555555557)
+('svm_comb_1', 15.88888888888889)
+('knn_50_comb_1', 16.77777777777778)
+('logit_comb_1', 17.22222222222222)
+20 20
+3.1253369501331285
+