机器学习-使用python进行谱聚类开发

import matplotlib.pyplot as plt import numpy as np from sklearn.cluster import SpectralClustering from sklearn.manifold import TSNE import umap.umap_ as umap import seaborn as sns from scipy.cluster.hierarchy import dendrogram, linkage from sklearn.datasets import load_iris # 加载Sklearn自带的数据集（鸢尾花数据集） iris = load_iris() X = iris.data y = iris.target # 使用谱聚类算法 n_clusters = 3 # 选择聚类数目 clustering = SpectralClustering(n_clusters=n_clusters, affinity='nearest_neighbors', random_state=42).fit(X) cluster_labels = clustering.labels_ # 计算相似性矩阵 # 根据聚类结果对数据进行排序 sorted_X = X[np.argsort(cluster_labels)] # 计算相似性矩阵 similarity_matrix = np.corrcoef(sorted_X) # 绘制相似性矩阵 plt.figure(figsize=(8, 6)) sns.heatmap(similarity_matrix, cmap='viridis') plt.title('Similarity Matrix') plt.xlabel('Samples') plt.ylabel('Samples') plt.savefig('./PNG/Similarity_Matrix.png') # 使用T-SNE降维可视化 tsne = TSNE(n_components=2, random_state=42) X_tsne = tsne.fit_transform(X) plt.figure(figsize=(8, 6)) plt.scatter(X_tsne[:, 0], X_tsne[:, 1], c=cluster_labels, cmap='viridis') plt.title('T-SNE') plt.xlabel('Component 1') plt.ylabel('Component 2') plt.colorbar(label='Cluster') plt.savefig('./PNG/T-SNE.png') # 使用UMAP降维可视化 umap_embedder = umap.UMAP(random_state=42) X_umap = umap_embedder.fit_transform(X) plt.figure(figsize=(8, 6)) plt.scatter(X_umap[:, 0], X_umap[:, 1], c=cluster_labels, cmap='viridis') plt.title('UMAP') plt.xlabel('UMAP 1') plt.ylabel('UMAP 2') plt.colorbar(label='Cluster') plt.savefig('./PNG/UMAP.png') # 绘制树状图 linkage_matrix = linkage(X, method='ward') plt.figure(figsize=(10, 8)) dendrogram(linkage_matrix) plt.title('Hierarchical Clustering Dendrogram') plt.xlabel('Sample Index') plt.ylabel('Distance') plt.savefig('./PNG/Hierarchical_Clustering_Dendrogram.png') # 绘制一致性热力图 plt.figure(figsize=(8, 6)) sns.heatmap(linkage_matrix, cmap='viridis', annot=True, fmt=".2f") plt.title('Consistency Heatmap') plt.xlabel('Sample Index') plt.ylabel('Sample Index') plt.savefig('./PNG/Consistency_Heatmap.png') plt.show() #把前面的plt.show()注释掉，然后运行这段代码，就可以实现同时产生很多图片，而不是关一个开下一个