Silohuette After Optimal Num of Clusters K-Means

app.py

@@ -22,6 +22,7 @@ import matplotlib.pyplot as plt
 import matplotlib.colors as mcolors
 import zipfile
 import tempfile
+from kneed import KneeLocator
 
 
 class RelativeScaler(BaseEstimator, TransformerMixin):
@@ -624,6 +625,13 @@ def compute_global_regression(df_combined, embedding_cols, tsne_params, df_f1, r
         inertias.append(kmeans.inertia_)
 
 
+    kl = KneeLocator(K, inertias, curve="convex", direction="decreasing")
+    elbow_k = kl.elbow
+    kmeans_opt = KMeans(n_clusters=elbow_k, random_state=42, n_init=10)
+    labels_opt = kmeans_opt.fit_predict(X)
+    silhouette_opt = silhouette_score(X, labels_opt)
+
+
     dfs_reduced, unique_subsets = split_versions(df_combined, reduced)
     
     df_distances = compute_cluster_distances_synthetic_individual(
@@ -700,6 +708,7 @@ def compute_global_regression(df_combined, embedding_cols, tsne_params, df_f1, r
         "continuity": cont,
         "silhouette": silhouette,
         "inertias": inertias,
+        "silhouette_opt": silhouette_opt,
     }
     
     if reduction_method == "PCA":
@@ -858,6 +867,8 @@ def run_model(model_name):
 
         st.bokeh_chart(p, use_container_width=True)
 
+    st.write(f"Silhouette Score: {result['silhouette_opt']:.2f}")
+
     # # Mostrar los plots de loadings si se usó PCA (para el conjunto combinado)
     # if reduction_method == "PCA" and result.get("pca_model") is not None:
     #     # pca_model = result["pca_model"]