Upload New File

tp_cluster.py

+import numpy as np
+import pandas as pd
+from sklearn.preprocessing import StandardScaler
+import scipy.cluster.hierarchy as hrch
+import matplotlib.pyplot as plt
+import sklearn.metrics as metr
+from sklearn.decomposition import PCA
+from sklearn.cluster import KMeans
+
+df=pd.read_csv("C:/Users/Antoine/Downloads/hotels.csv")
+
+#Corrélation des données
+print(df.iloc[:,2:].corr())
+
+
+#On centre et réduit les données
+SS= StandardScaler()
+
+df_centre_reduit=SS.fit_transform(df[list(df)[2:]])
+
+#mean~=0 et std=1 Tout est ok
+print(np.mean(df_centre_reduit,axis=0))
+print(np.std(df_centre_reduit,axis=0))
+
+# Arborescence hiérarchique + affichage de l'arbre
+L=hrch.linkage(df_centre_reduit)
+# hrch.dendrogram(L)
+# plt.show()
+
+# FC Contiendra les clusters
+FC=hrch.fcluster(L,2.1,criterion="distance")
+print(metr.silhouette_score(df_centre_reduit,FC))
+
+#PCA pour visualisation
+my_PCA=PCA()
+pca_df=my_PCA.fit_transform(df_centre_reduit)
+
+
+#Couleurs pour chaque cluster
+couleurs=["red","blue","green","yellow","orange","purple","black","gray","rose"]
+c=[couleurs[i-1] for i in FC]
+plt.scatter(pca_df[:,0],pca_df[:,1],c=c)
+# plt.show()
+
+
+## KMeans n_cluster à changer
+n_cluster=5
+my_kmeans=KMeans(n_cluster)
+
+my_kmeans.fit_transform(df_centre_reduit)
+
+c=[couleurs[i] for i in my_kmeans.labels_]
+plt.scatter(pca_df[:,0],pca_df[:,1],c=c)
+# plt.show()
\ No newline at end of file