mc

TASK_2/silhouette_random_labels.py

@@ -1,3 +1,8 @@
+# %% [markdown]
+# **Data mining Project - University of Pisa, acedemic year 2023/24**
+#
+# **Authors**: Giacomo Aru, Giulia Ghisolfi, Luca Marini, Irene Testa
+
 # %%
 import json
 import pandas as pd
@@ -29,7 +34,7 @@
 for n_cluster in n_clusters:
     print(f'Trying {n_cluster} clusters')
     silhouette_values = []
-    for i in range(10): # TODO: aumentare? con 10 iterazioni per una size ci mette  15 min
+    for i in range(10):
         print(f'Iteration {i}')
         labels = np.random.randint(0, n_cluster, len(indicators_df))
         silhouette_values.append(silhouette_score(X, labels))

TASK_4/shap_giacomo.ipynb

@@ -43,7 +43,7 @@
     "RANDOM_STATE = 42\n",
     "NSAMPLE = 1000\n",
     "NBASETRAINISTANCES = 20\n",
-    "SVMISTANCESTOCOMPUTE = 150 # deve essere minore di NSAMPLE e maggiore di 66\n",
+    "SVMISTANCESTOCOMPUTE = 150\n",
     "\n",
     "pd.set_option('display.max_columns', None)\n",
     "pd.set_option('max_colwidth', None)\n",
@@ -107,8 +107,6 @@
     "true_labels_test_df = pd.read_csv('../data/clf_y_test.csv', index_col=0)\n",
     "true_labels_test = true_labels_test_df.values.ravel()\n",
     "\n",
-    "\n",
-    "\n",
     "# load the names of the features\n",
     "features_db = json.loads(open('../data/clf_indicators_names_distance_based.json').read())\n",
     "features_rb = json.loads(open('../data/clf_indicators_names_rule_based.json').read())\n",

index.md

@@ -1,6 +1,3 @@
-TODO:
-- fare il rendering dei notebook e linkarli qui
-
 ## Data visualization
 This repository hosts also interactive data visualizations. You can access and interact with the plots by clicking the links below.