{"id":1032967,"date":"2024-12-31T11:37:59","date_gmt":"2024-12-31T03:37:59","guid":{"rendered":"https:\/\/docs.pingcode.com\/ask\/ask-ask\/1032967.html"},"modified":"2024-12-31T11:38:01","modified_gmt":"2024-12-31T03:38:01","slug":"python%e5%a6%82%e4%bd%95%e5%af%b9%e7%ba%bf%e6%ae%b5%e8%bf%9b%e8%a1%8c%e8%bd%a8%e8%bf%b9%e8%81%9a%e7%b1%bb","status":"publish","type":"post","link":"https:\/\/docs.pingcode.com\/ask\/1032967.html","title":{"rendered":"python\u5982\u4f55\u5bf9\u7ebf\u6bb5\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b"},"content":{"rendered":"

\"python\u5982\u4f55\u5bf9\u7ebf\u6bb5\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b\"<\/p>\n

Python\u53ef\u4ee5\u901a\u8fc7\u4f7f\u7528DBSCAN\u7b97\u6cd5\u3001OPTICS\u7b97\u6cd5\u3001Mean Shift\u7b97\u6cd5\u6765\u5bf9\u7ebf\u6bb5\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b\u3002DBSCAN\u7b97\u6cd5\u662f\u5bc6\u5ea6\u805a\u7c7b\u7b97\u6cd5\uff0c\u53ef\u4ee5\u8bc6\u522b\u4efb\u610f\u5f62\u72b6\u7684\u805a\u7c7b\uff0c\u9002\u5408\u5904\u7406\u5e26\u6709\u566a\u58f0\u7684\u6570\u636e\u3002<\/strong><\/p>\n<\/p>\n

\u5177\u4f53\u6765\u8bf4\uff0cDBSCAN\u7b97\u6cd5\u901a\u8fc7\u6307\u5b9a\u534a\u5f84\u53c2\u6570\u548c\u6700\u5c0f\u6837\u672c\u6570\u6765\u5b9a\u4e49\u7c07\u7684\u5bc6\u5ea6\uff0c\u4ece\u800c\u5c06\u8f68\u8ff9\u805a\u7c7b\u5230\u4e00\u8d77\u3002\u6211\u4eec\u53ef\u4ee5\u5c06\u8f68\u8ff9\u7684\u8d77\u70b9\u548c\u7ec8\u70b9\u4f5c\u4e3a\u6570\u636e\u70b9\u8f93\u5165\u5230DBSCAN\u7b97\u6cd5\u4e2d\uff0c\u7b97\u6cd5\u4f1a\u6839\u636e\u70b9\u7684\u5bc6\u5ea6\u5c06\u5176\u805a\u7c7b\u3002<\/p>\n<\/p>\n

\u4e00\u3001\u5b89\u88c5\u548c\u5bfc\u5165\u5fc5\u8981\u7684\u5e93<\/h3>\n<\/p>\n

import numpy as np<\/p>\n
import matplotlib.pyplot as plt<\/p>\n
from sklearn.cluster import DBSCAN<\/p>\n
from sklearn.preprocessing import StandardScaler<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u4f7f\u7528DBSCAN\u7b97\u6cd5\u4e4b\u524d\uff0c\u6211\u4eec\u9996\u5148\u9700\u8981\u5b89\u88c5\u5e76\u5bfc\u5165\u5fc5\u8981\u7684\u5e93\u3002numpy<\/code>\u5e93\u7528\u4e8e\u6570\u503c\u8ba1\u7b97\uff0cmatplotlib<\/code>\u5e93\u7528\u4e8e\u6570\u636e\u53ef\u89c6\u5316\uff0csklearn<\/code>\u5e93\u4e2d\u7684DBSCAN<\/code>\u7c7b\u7528\u4e8e\u8fdb\u884c\u5bc6\u5ea6\u805a\u7c7b\uff0cStandardScaler<\/code>\u7c7b\u7528\u4e8e\u6570\u636e\u6807\u51c6\u5316\u3002<\/p>\n<\/p>\n
\u4e8c\u3001\u751f\u6210\u793a\u4f8b\u6570\u636e<\/h3>\n<\/p>\n
# \u751f\u6210\u793a\u4f8b\u6570\u636e\uff1a\u7ebf\u6bb5\u7684\u8d77\u70b9\u548c\u7ec8\u70b9<\/p>\n
line_segments = np.array([<\/p>\n
    [0, 0, 1, 1],<\/p>\n
    [1, 1, 2, 2],<\/p>\n
    [2, 2, 3, 3],<\/p>\n
    [8, 8, 9, 9],<\/p>\n
    [9, 9, 10, 10],<\/p>\n
    [10, 10, 11, 11]<\/p>\n
])<\/p>\n
\u63d0\u53d6\u7ebf\u6bb5\u7684\u8d77\u70b9\u548c\u7ec8\u70b9<\/strong><\/h2>\n
points = np.vstack((line_segments[:, :2], line_segments[:, 2:]))<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e00\u6b65\uff0c\u6211\u4eec\u751f\u6210\u4e86\u4e00\u4e9b\u793a\u4f8b\u6570\u636e\uff0c\u8868\u793a\u4e00\u4e9b\u7ebf\u6bb5\u7684\u8d77\u70b9\u548c\u7ec8\u70b9\u3002\u4e3a\u4e86\u65b9\u4fbf\u805a\u7c7b\uff0c\u6211\u4eec\u5c06\u8fd9\u4e9b\u8d77\u70b9\u548c\u7ec8\u70b9\u63d0\u53d6\u51fa\u6765\u5e76\u7ec4\u5408\u6210\u4e00\u4e2a\u70b9\u96c6\u3002<\/p>\n<\/p>\n
\u4e09\u3001\u6570\u636e\u6807\u51c6\u5316<\/h3>\n<\/p>\n
scaler = StandardScaler()<\/p>\n
points_scaled = scaler.fit_transform(points)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u4e3a\u4e86\u66f4\u597d\u5730\u8fdb\u884c\u805a\u7c7b\uff0c\u6211\u4eec\u9700\u8981\u5bf9\u6570\u636e\u8fdb\u884c\u6807\u51c6\u5316\u5904\u7406\uff0c\u4f7f\u5176\u7b26\u5408\u6807\u51c6\u6b63\u6001\u5206\u5e03\u3002\u6211\u4eec\u4f7f\u7528StandardScaler<\/code>\u7c7b\u6765\u5b8c\u6210\u8fd9\u4e00\u64cd\u4f5c\u3002<\/p>\n<\/p>\n
\u56db\u3001\u5e94\u7528DBSCAN\u7b97\u6cd5\u8fdb\u884c\u805a\u7c7b<\/h3>\n<\/p>\n
db = DBSCAN(eps=0.5, min_samples=2).fit(points_scaled)<\/p>\n
labels = db.labels_<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e00\u6b65\uff0c\u6211\u4eec\u4f7f\u7528DBSCAN\u7b97\u6cd5\u5bf9\u6807\u51c6\u5316\u540e\u7684\u6570\u636e\u8fdb\u884c\u805a\u7c7b\u3002eps<\/code>\u53c2\u6570\u8868\u793a\u90bb\u57df\u7684\u534a\u5f84\uff0cmin_samples<\/code>\u53c2\u6570\u8868\u793a\u5b9a\u4e49\u4e00\u4e2a\u7c07\u7684\u6700\u5c0f\u6837\u672c\u6570\u3002\u7b97\u6cd5\u4f1a\u6839\u636e\u70b9\u7684\u5bc6\u5ea6\u5c06\u5176\u805a\u7c7b\uff0c\u5e76\u751f\u6210\u6bcf\u4e2a\u70b9\u7684\u7c07\u6807\u7b7e\u3002<\/p>\n<\/p>\n
\u4e94\u3001\u7ed3\u679c\u53ef\u89c6\u5316<\/h3>\n<\/p>\n
# \u6839\u636e\u805a\u7c7b\u7ed3\u679c\u7ed8\u5236\u7ebf\u6bb5<\/p>\n
unique_labels = set(labels)<\/p>\n
colors = [plt.cm.Spectral(each) for each in np.linspace(0, 1, len(unique_labels))]<\/p>\n
for k, col in zip(unique_labels, colors):<\/p>\n
    if k == -1:<\/p>\n
        # \u566a\u58f0\u70b9<\/p>\n
        col = [0, 0, 0, 1]<\/p>\n
    class_member_mask = (labels == k)<\/p>\n
    xy = points[class_member_mask]<\/p>\n
    plt.plot(xy[:, 0], xy[:, 1], 'o', markerfacecolor=tuple(col), markeredgecolor='k', markersize=14)<\/p>\n
plt.title('DBSCAN Clustering of Line Segments')<\/p>\n
plt.show()<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u6700\u540e\uff0c\u6211\u4eec\u6839\u636e\u805a\u7c7b\u7ed3\u679c\u7ed8\u5236\u7ebf\u6bb5\u3002\u6211\u4eec\u4e3a\u4e0d\u540c\u7684\u7c07\u4f7f\u7528\u4e0d\u540c\u7684\u989c\u8272\u8fdb\u884c\u6807\u8bc6\uff0c\u5e76\u5c06\u566a\u58f0\u70b9\uff08\u6807\u7b7e\u4e3a-1\u7684\u70b9\uff09\u6807\u8bb0\u4e3a\u9ed1\u8272\u3002\u901a\u8fc7\u53ef\u89c6\u5316\uff0c\u6211\u4eec\u53ef\u4ee5\u76f4\u89c2\u5730\u67e5\u770b\u805a\u7c7b\u7ed3\u679c\u3002<\/p>\n<\/p>\n
\u516d\u3001\u6269\u5c55\uff1a\u4f7f\u7528OPTICS\u7b97\u6cd5\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b<\/h3>\n<\/p>\n
\u9664\u4e86DBSCAN\u7b97\u6cd5\uff0cOPTICS\u7b97\u6cd5\u4e5f\u662f\u4e00\u79cd\u5e38\u7528\u7684\u5bc6\u5ea6\u805a\u7c7b\u7b97\u6cd5\uff0c\u7279\u522b\u9002\u7528\u4e8e\u5904\u7406\u5177\u6709\u4e0d\u540c\u5bc6\u5ea6\u7684\u7c07\u3002\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528sklearn<\/code>\u5e93\u4e2d\u7684OPTICS<\/code>\u7c7b\u6765\u5b9e\u73b0\u8f68\u8ff9\u805a\u7c7b\u3002<\/p>\n<\/p>\n
from sklearn.cluster import OPTICS<\/p>\n
\u4f7f\u7528OPTICS\u7b97\u6cd5\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b<\/strong><\/h2>\n
optics = OPTICS(min_samples=2).fit(points_scaled)<\/p>\n
optics_labels = optics.labels_<\/p>\n
\u6839\u636e\u805a\u7c7b\u7ed3\u679c\u7ed8\u5236\u7ebf\u6bb5<\/strong><\/h2>\n
unique_labels = set(optics_labels)<\/p>\n
colors = [plt.cm.Spectral(each) for each in np.linspace(0, 1, len(unique_labels))]<\/p>\n
for k, col in zip(unique_labels, colors):<\/p>\n
    if k == -1:<\/p>\n
        # \u566a\u58f0\u70b9<\/p>\n
        col = [0, 0, 0, 1]<\/p>\n
    class_member_mask = (optics_labels == k)<\/p>\n
    xy = points[class_member_mask]<\/p>\n
    plt.plot(xy[:, 0], xy[:, 1], 'o', markerfacecolor=tuple(col), markeredgecolor='k', markersize=14)<\/p>\n
plt.title('OPTICS Clustering of Line Segments')<\/p>\n
plt.show()<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u6bb5\u4ee3\u7801\u4e2d\uff0c\u6211\u4eec\u4f7f\u7528OPTICS\u7b97\u6cd5\u5bf9\u6807\u51c6\u5316\u540e\u7684\u6570\u636e\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b\uff0c\u5e76\u7ed8\u5236\u805a\u7c7b\u7ed3\u679c\u3002OPTICS\u7b97\u6cd5\u4e0d\u9700\u8981\u6307\u5b9a\u90bb\u57df\u534a\u5f84\u53c2\u6570\uff0c\u9002\u7528\u4e8e\u5904\u7406\u5177\u6709\u4e0d\u540c\u5bc6\u5ea6\u7684\u7c07\u3002<\/p>\n<\/p>\n
\u4e03\u3001\u6269\u5c55\uff1a\u4f7f\u7528Mean Shift\u7b97\u6cd5\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b<\/h3>\n<\/p>\n
Mean Shift\u7b97\u6cd5\u662f\u4e00\u79cd\u57fa\u4e8e\u5bc6\u5ea6\u7684\u805a\u7c7b\u65b9\u6cd5\uff0c\u901a\u8fc7\u5bfb\u627e\u6570\u636e\u70b9\u5bc6\u5ea6\u7684\u5cf0\u503c\u6765\u5b9a\u4e49\u7c07\u3002\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528sklearn<\/code>\u5e93\u4e2d\u7684MeanShift<\/code>\u7c7b\u6765\u5b9e\u73b0\u8f68\u8ff9\u805a\u7c7b\u3002<\/p>\n<\/p>\n
from sklearn.cluster import MeanShift<\/p>\n
\u4f7f\u7528Mean Shift\u7b97\u6cd5\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b<\/strong><\/h2>\n
mean_shift = MeanShift().fit(points_scaled)<\/p>\n
mean_shift_labels = mean_shift.labels_<\/p>\n
\u6839\u636e\u805a\u7c7b\u7ed3\u679c\u7ed8\u5236\u7ebf\u6bb5<\/strong><\/h2>\n
unique_labels = set(mean_shift_labels)<\/p>\n
colors = [plt.cm.Spectral(each) for each in np.linspace(0, 1, len(unique_labels))]<\/p>\n
for k, col in zip(unique_labels, colors):<\/p>\n
    if k == -1:<\/p>\n
        # \u566a\u58f0\u70b9<\/p>\n
        col = [0, 0, 0, 1]<\/p>\n
    class_member_mask = (mean_shift_labels == k)<\/p>\n
    xy = points[class_member_mask]<\/p>\n
    plt.plot(xy[:, 0], xy[:, 1], 'o', markerfacecolor=tuple(col), markeredgecolor='k', markersize=14)<\/p>\n
plt.title('Mean Shift Clustering of Line Segments')<\/p>\n
plt.show()<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u6bb5\u4ee3\u7801\u4e2d\uff0c\u6211\u4eec\u4f7f\u7528Mean Shift\u7b97\u6cd5\u5bf9\u6807\u51c6\u5316\u540e\u7684\u6570\u636e\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b\uff0c\u5e76\u7ed8\u5236\u805a\u7c7b\u7ed3\u679c\u3002Mean Shift\u7b97\u6cd5\u4e0d\u9700\u8981\u6307\u5b9a\u7c07\u7684\u6570\u91cf\uff0c\u9002\u7528\u4e8e\u5904\u7406\u5177\u6709\u4e0d\u540c\u5f62\u72b6\u548c\u5bc6\u5ea6\u7684\u7c07\u3002<\/p>\n<\/p>\n
\u603b\u7ed3<\/h3>\n<\/p>\n
\u901a\u8fc7\u4f7f\u7528DBSCAN\u3001OPTICS\u548cMean Shift\u7b49\u5bc6\u5ea6\u805a\u7c7b\u7b97\u6cd5\uff0c\u6211\u4eec\u53ef\u4ee5\u65b9\u4fbf\u5730\u5bf9\u7ebf\u6bb5\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b\u3002DBSCAN\u7b97\u6cd5\u9002\u7528\u4e8e\u8bc6\u522b\u4efb\u610f\u5f62\u72b6\u7684\u805a\u7c7b\uff0c\u5e76\u4e14\u53ef\u4ee5\u5904\u7406\u5e26\u6709\u566a\u58f0\u7684\u6570\u636e\u3002OPTICS\u7b97\u6cd5\u9002\u7528\u4e8e\u5904\u7406\u5177\u6709\u4e0d\u540c\u5bc6\u5ea6\u7684\u7c07\uff0c\u800cMean Shift\u7b97\u6cd5\u9002\u7528\u4e8e\u5904\u7406\u5177\u6709\u4e0d\u540c\u5f62\u72b6\u548c\u5bc6\u5ea6\u7684\u7c07\u3002\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\uff0c\u6211\u4eec\u53ef\u4ee5\u6839\u636e\u6570\u636e\u7684\u7279\u70b9\u9009\u62e9\u5408\u9002\u7684\u805a\u7c7b\u7b97\u6cd5\uff0c\u4ee5\u83b7\u5f97\u6700\u4f73\u7684\u805a\u7c7b\u6548\u679c\u3002<\/p>\n<\/p>\n
\u76f8\u5173\u95ee\u7b54FAQs\uff1a<\/strong><\/h2>\n
 \u5982\u4f55\u9009\u62e9\u9002\u5408\u7684\u8f68\u8ff9\u805a\u7c7b\u7b97\u6cd5\uff1f<\/strong>
\u5728\u9009\u62e9\u8f68\u8ff9\u805a\u7c7b\u7b97\u6cd5\u65f6\uff0c\u9700\u8981\u8003\u8651\u6570\u636e\u7684\u7279\u6027\u548c\u5e94\u7528\u573a\u666f\u3002\u5e38\u89c1\u7684\u7b97\u6cd5\u6709DBSCAN\u3001K-means\u548cOPTICS\u7b49\u3002DBSCAN\u9002\u5408\u5904\u7406\u566a\u58f0\u548c\u4e0d\u89c4\u5219\u5f62\u72b6\u7684\u8f68\u8ff9\uff0c\u800cK-means\u9002\u5408\u5904\u7406\u5df2\u77e5\u6570\u91cf\u7684\u8f68\u8ff9\u805a\u7c7b\u3002\u4e86\u89e3\u6570\u636e\u7684\u5206\u5e03\u60c5\u51b5\u548c\u805a\u7c7b\u76ee\u6807\u53ef\u4ee5\u5e2e\u52a9\u4f60\u505a\u51fa\u66f4\u5408\u9002\u7684\u9009\u62e9\u3002<\/p>\n
Python\u4e2d\u6709\u54ea\u4e9b\u5e93\u53ef\u4ee5\u5b9e\u73b0\u8f68\u8ff9\u805a\u7c7b\uff1f<\/strong>
Python\u4e2d\u6709\u591a\u4e2a\u5e93\u53ef\u4ee5\u5e2e\u52a9\u5b9e\u73b0\u8f68\u8ff9\u805a\u7c7b\uff0c\u5e38\u7528\u7684\u5305\u62ecScikit-learn\u3001HDBSCAN\u548cPyClustering\u3002Scikit-learn\u63d0\u4f9b\u591a\u79cd\u805a\u7c7b\u7b97\u6cd5\u7684\u5b9e\u73b0\uff0cHDBSCAN\u5219\u4e13\u6ce8\u4e8e\u5904\u7406\u9ad8\u7ef4\u6570\u636e\u7684\u805a\u7c7b\u95ee\u9898\uff0cPyClustering\u5219\u63d0\u4f9b\u591a\u79cd\u805a\u7c7b\u65b9\u6cd5\u548c\u76f8\u5173\u5de5\u5177\uff0c\u9002\u5408\u591a\u79cd\u5e94\u7528\u9700\u6c42\u3002<\/p>\n
\u5982\u4f55\u8bc4\u4f30\u8f68\u8ff9\u805a\u7c7b\u7684\u6548\u679c\uff1f<\/strong>
\u8bc4\u4f30\u8f68\u8ff9\u805a\u7c7b\u6548\u679c\u7684\u65b9\u6cd5\u6709\u591a\u79cd\uff0c\u5e38\u7528\u7684\u5305\u62ec\u8f6e\u5ed3\u7cfb\u6570\u3001Calinski-Harabasz\u6307\u6570\u548cDavies-Bouldin\u6307\u6570\u7b49\u3002\u8f6e\u5ed3\u7cfb\u6570\u53ef\u4ee5\u8861\u91cf\u805a\u7c7b\u7684\u7d27\u5bc6\u5ea6\u548c\u5206\u79bb\u5ea6\uff0cCalinski-Harabasz\u6307\u6570\u5219\u901a\u8fc7\u6bd4\u8f83\u7c07\u5185\u548c\u7c07\u95f4\u7684\u65b9\u5dee\u6765\u8bc4\u4f30\u805a\u7c7b\u6548\u679c\uff0c\u800cDavies-Bouldin\u6307\u6570\u5219\u901a\u8fc7\u7c07\u95f4\u8ddd\u79bb\u548c\u7c07\u5185\u8ddd\u79bb\u7684\u6bd4\u7387\u6765\u8fdb\u884c\u8bc4\u4f30\u3002\u9009\u62e9\u5408\u9002\u7684\u8bc4\u4f30\u6307\u6807\u53ef\u4ee5\u5e2e\u52a9\u4f60\u66f4\u597d\u5730\u7406\u89e3\u805a\u7c7b\u7ed3\u679c\u7684\u8d28\u91cf\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"Python\u53ef\u4ee5\u901a\u8fc7\u4f7f\u7528DBSCAN\u7b97\u6cd5\u3001OPTICS\u7b97\u6cd5\u3001Mean Shift\u7b97\u6cd5\u6765\u5bf9\u7ebf\u6bb5\u8fdb\u884c\u8f68\u8ff9\u805a\u7c7b\u3002DB […]","protected":false},"author":3,"featured_media":1032975,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[37],"tags":[],"acf":[],"_links":{"self":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/1032967"}],"collection":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/comments?post=1032967"}],"version-history":[{"count":"1","href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/1032967\/revisions"}],"predecessor-version":[{"id":1032976,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/1032967\/revisions\/1032976"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media\/1032975"}],"wp:attachment":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media?parent=1032967"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/categories?post=1032967"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/tags?post=1032967"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}