{"id":2778,"date":"2024-10-26T07:56:22","date_gmt":"2024-10-25T23:56:22","guid":{"rendered":"http:\/\/viplao.com\/?p=2778"},"modified":"2024-10-27T22:33:49","modified_gmt":"2024-10-27T14:33:49","slug":"%e6%95%b0%e6%8d%ae%e5%88%86%e6%9e%90-%e5%ae%a2%e6%88%b7%e7%89%b9%e5%be%81%e7%9a%84%e8%81%9a%e7%b1%bb%e5%88%86%e6%9e%90%e5%ae%9e%e8%b7%b5%e6%a1%88%e4%be%8b","status":"publish","type":"post","link":"http:\/\/viplao.com\/index.php\/2024\/10\/26\/%e6%95%b0%e6%8d%ae%e5%88%86%e6%9e%90-%e5%ae%a2%e6%88%b7%e7%89%b9%e5%be%81%e7%9a%84%e8%81%9a%e7%b1%bb%e5%88%86%e6%9e%90%e5%ae%9e%e8%b7%b5%e6%a1%88%e4%be%8b\/","title":{"rendered":"\u6570\u5b57\u5316\u8fd0\u8425\u57fa\u7840\u6280\u80fd &#8211; \u5ba2\u6237\u7279\u5f81\u7684\u805a\u7c7b\u5206\u6790\u5b9e\u8df5\u6848\u4f8b"},"content":{"rendered":"\n<pre class=\"wp-block-code\"><code># \u5bfc\u5165\u5e93\nimport pandas as pd # panda\u5e93\nimport numpy as np\nimport matplotlib.pyplot as plt  # \u5bfc\u5165matplotlib\u5e93\nfrom sklearn.preprocessing import MinMaxScaler # \u6807\u51c6\u5316\u5e93\nfrom sklearn.cluster import KMeans  # \u5bfc\u5165sklearn\u805a\u7c7b\u6a21\u5757\nfrom sklearn.metrics import silhouette_score   # \u6548\u679c\u8bc4\u4f30\u6a21\u5757\uff0c \u65b0\u7248\u672c\u4e2d\u5df2\u7ecf\u6ca1\u6709 calinski_harabaz_score \u65b9\u6cd5\nimport matplotlib.pyplot as plt # \u56fe\u5f62\u5e93<\/code><\/pre>\n\n\n\n<p># \u8bfb\u53d6\u6570\u636e<\/p>\n\n\n\n<p>raw_data = pd.read_csv(&#8216;cluster.txt&#8217;) &nbsp;# \u5bfc\u5165\u6570\u636e\u6587\u4ef6<\/p>\n\n\n\n<p>numeric_features = raw_data.iloc[:,1:3] # \u6570\u503c\u578b\u7279\u5f81<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># \u6570\u636e\u6807\u51c6\u5316\nscaler = MinMaxScaler()\nscaled_numeric_features = scaler.fit_transform(numeric_features)\nprint(scaled_numeric_features&#91;:,:2])<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code># \u8bad\u7ec3\u805a\u7c7b\u6a21\u578b\nn_clusters = 3  # \u8bbe\u7f6e\u805a\u7c7b\u6570\u91cf\nmodel_kmeans = KMeans(n_clusters=n_clusters, random_state=0)  # \u5efa\u7acb\u805a\u7c7b\u6a21\u578b\u5bf9\u8c61\nmodel_kmeans.fit(scaled_numeric_features)  # \u8bad\u7ec3\u805a\u7c7b\u6a21\u578b<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code># \u6a21\u578b\u6548\u679c\u6307\u6807\u8bc4\u4f30\n# \u603b\u6837\u672c\u91cf,\u603b\u7279\u5f81\u6570\nn_samples, n_features = raw_data.iloc&#91;:,1:].shape\nprint('samples: %d \\t features: %d' % (n_samples, n_features))\n\n# \u975e\u76d1\u7763\u5f0f\u8bc4\u4f30\u65b9\u6cd5\nsilhouette_s = silhouette_score(scaled_numeric_features, model_kmeans.labels_, metric='euclidean')  # \u5e73\u5747\u8f6e\u5ed3\u7cfb\u6570\n# calinski_harabaz_s = calinski_harabaz_score(scaled_numeric_features, model_kmeans.labels_)  # \u8001\u7248\u672c\u6709\uff0c\u65b0\u7248\u672c\u6ca1\u6709\u8be5\u65b9\u6cd5\u4e86\n# unsupervised_data = {'silh':&#91;silhouette_s],'c&amp;h':&#91;calinski_harabaz_s]} # \u8001\u7248\u672c\u65b9\u6cd5\nunsupervised_data = {'silh':&#91;silhouette_s]} # \u65b0\u7248\u672c\u65b9\u6cd5\nunsupervised_score = pd.DataFrame.from_dict(unsupervised_data)\nprint('\\n','unsupervised score:','\\n','-'*60)\nprint(unsupervised_score)<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code># \u5408\u5e76\u6570\u636e\u548c\u7279\u5f81\n# \u83b7\u5f97\u6bcf\u4e2a\u6837\u672c\u7684\u805a\u7c7b\u7c7b\u522b\nkmeans_labels = pd.DataFrame(model_kmeans.labels_,columns=&#91;'labels']) \n# \u7ec4\u5408\u539f\u59cb\u6570\u636e\u4e0e\u6807\u7b7e\nkmeans_data = pd.concat((raw_data,kmeans_labels),axis=1)\nprint(kmeans_data.head())<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code># \u8ba1\u7b97\u4e0d\u540c\u805a\u7c7b\u7c7b\u522b\u7684\u6837\u672c\u91cf\u548c\u5360\u6bd4\nlabel_count = kmeans_data.groupby(&#91;'labels'])&#91;'SEX'].count()  # \u8ba1\u7b97\u9891\u6570\nlabel_count_rate = label_count\/ kmeans_data.shape&#91;0] # \u8ba1\u7b97\u5360\u6bd4\nkmeans_record_count = pd.concat((label_count,label_count_rate),axis=1)\nkmeans_record_count.columns=&#91;'record_count','record_rate']\nprint(kmeans_record_count.head())<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code># \u8ba1\u7b97\u4e0d\u540c\u805a\u7c7b\u7c7b\u522b\u6570\u503c\u578b\u7279\u5f81 'AVG_ORDERS',\nkmeans_numeric_features = kmeans_data.groupby(&#91;'labels'])&#91;'AVG_MONEY'].mean()\n\nkmeans_avg_orders = kmeans_data.groupby(&#91;'labels'])&#91;'AVG_ORDERS'].mean()\nprint(kmeans_numeric_features)\nprint(kmeans_numeric_features.head())<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code># \u8ba1\u7b97\u4e0d\u540c\u805a\u7c7b\u7c7b\u522b\u5206\u7c7b\u578b\u7279\u5f81\nactive_list = &#91;]\nsex_gb_list = &#91;]\nunique_labels = np.unique(model_kmeans.labels_)\nfor each_label in unique_labels:\n    each_data = kmeans_data&#91;kmeans_data&#91;'labels']==each_label]\n    active_list.append(each_data.groupby(&#91;'IS_ACTIVE'])&#91;'USER_ID'].count()\/each_data.shape&#91;0])\n    sex_gb_list.append(each_data.groupby(&#91;'SEX'])&#91;'USER_ID'].count()\/each_data.shape&#91;0])\n\nkmeans_active_pd = pd.DataFrame(active_list)\nkmeans_sex_gb_pd = pd.DataFrame(sex_gb_list)\nkmeans_string_features = pd.concat((kmeans_active_pd,kmeans_sex_gb_pd),axis=1)\nkmeans_string_features.index = unique_labels\nprint(kmeans_string_features.head())<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code>features_all2 = pd.concat((kmeans_record_count,kmeans_numeric_features,kmeans_avg_orders,kmeans_string_features),axis=1)\nprint(features_all2.head())<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code># \u5408\u5e76\u6240\u6709\u7c7b\u522b\u7684\u5206\u6790\u7ed3\u679c\nfeatures_all =  pd.concat((kmeans_record_count,kmeans_numeric_features,kmeans_avg_orders,kmeans_string_features),axis=1)\nprint(features_all.head())<\/code><\/pre>\n\n\n\n<pre class=\"wp-block-code\"><code># \u53ef\u89c6\u5316\u56fe\u5f62\u5c55\u793a\n# part 1 \u5168\u5c40\u914d\u7f6e\nfig = plt.figure(figsize=(10, 7))\ntitles = &#91;'RECORD_RATE','AVG_ORDERS','AVG_MONEY','IS_ACTIVE','SEX'] # \u5171\u7528\u6807\u9898\nline_index,col_index = 3,5 # \u5b9a\u4e49\u7f51\u683c\u6570\nax_ids = np.arange(1,16).reshape(line_index,col_index) # \u751f\u6210\u5b50\u7f51\u683c\u7d22\u5f15\u503c\nplt.rcParams&#91;'font.sans-serif']=&#91;'SimHei'] #\u7528\u6765\u6b63\u5e38\u663e\u793a\u4e2d\u6587\u6807\u7b7e\n    \n# part 2 \u753b\u51fa\u4e09\u4e2a\u7c7b\u522b\u7684\u5360\u6bd4\npie_fracs = features_all&#91;'record_rate'].tolist()\nfor ind in range(len(pie_fracs)):\n    ax = fig.add_subplot(line_index, col_index, ax_ids&#91;:,0]&#91;ind])\n    init_labels = &#91;'','',''] # \u521d\u59cb\u5316\u7a7alabel\u6807\u7b7e\n    init_labels&#91;ind] = 'cluster_{0}'.format(ind) # \u8bbe\u7f6e\u6807\u7b7e\n    init_colors = &#91;'lightgray', 'lightgray', 'lightgray']\n    init_colors&#91;ind] = 'g' # \u8bbe\u7f6e\u76ee\u6807\u9762\u79ef\u533a\u522b\u989c\u8272\n    ax.pie(x=pie_fracs, autopct='%3.0f %%',labels=init_labels,colors=init_colors)\n    ax.set_aspect('equal') # \u8bbe\u7f6e\u997c\u56fe\u4e3a\u5706\u5f62\n    if ind == 0:\n        ax.set_title(titles&#91;0])\n    \n# part 3  \u753b\u51faAVG_ORDERS\u5747\u503c\navg_orders_label = 'AVG_ORDERS'\navg_orders_fraces = features_all&#91;avg_orders_label]\n\nfor ind, frace in enumerate(avg_orders_fraces):\n    ax = fig.add_subplot(line_index, col_index, ax_ids&#91;:,1]&#91;ind])\n    ax.bar(x=unique_labels,height=&#91;0,avg_orders_fraces&#91;ind],0])# \u753b\u51fa\u67f1\u5f62\u56fe\n    ax.set_ylim((0, max(avg_orders_fraces)*1.2))\n    ax.set_xticks(&#91;])\n    ax.set_yticks(&#91;])\n    if ind == 0:# \u8bbe\u7f6e\u603b\u6807\u9898\n        ax.set_title(titles&#91;1])\n    # \u8bbe\u7f6e\u6bcf\u4e2a\u67f1\u5f62\u56fe\u7684\u6570\u503c\u6807\u7b7e\u548cx\u8f74label\n    ax.text(unique_labels&#91;1],frace+0.4,s='{:.2f}'.format(frace),ha='center',va='top')\n    ax.text(unique_labels&#91;1],-0.4,s=avg_orders_label,ha='center',va='bottom')\n        \n# part 4  \u753b\u51faAVG_MONEY\u5747\u503c\navg_money_label = 'AVG_MONEY'\navg_money_fraces = features_all&#91;avg_money_label]\nfor ind, frace in enumerate(avg_money_fraces):\n    ax = fig.add_subplot(line_index, col_index, ax_ids&#91;:,2]&#91;ind])\n    ax.bar(x=unique_labels,height=&#91;0,avg_money_fraces&#91;ind],0])# \u753b\u51fa\u67f1\u5f62\u56fe\n    ax.set_ylim((0, max(avg_money_fraces)*1.2))\n    ax.set_xticks(&#91;])\n    ax.set_yticks(&#91;])\n    if ind == 0:# \u8bbe\u7f6e\u603b\u6807\u9898\n        ax.set_title(titles&#91;2])\n    # \u8bbe\u7f6e\u6bcf\u4e2a\u67f1\u5f62\u56fe\u7684\u6570\u503c\u6807\u7b7e\u548cx\u8f74label\n    ax.text(unique_labels&#91;1],frace+4,s='{:.0f}'.format(frace),ha='center',va='top')\n    ax.text(unique_labels&#91;1],-4,s=avg_money_label,ha='center',va='bottom')\n        \n# part 5  \u753b\u51fa\u662f\u5426\u6d3b\u8dc3\naxtivity_labels = &#91;'\u4e0d\u6d3b\u8dc3','\u6d3b\u8dc3']\nx_ticket = &#91;i for i in range(len(axtivity_labels))]\nactivity_data = features_all&#91;axtivity_labels]\nylim_max = np.max(np.max(activity_data))\nfor ind,each_data in enumerate(activity_data.values):\n    ax = fig.add_subplot(line_index, col_index, ax_ids&#91;:,3]&#91;ind])\n    ax.bar(x=x_ticket,height=each_data) # \u753b\u51fa\u67f1\u5f62\u56fe\n    ax.set_ylim((0, ylim_max*1.2))\n    ax.set_xticks(&#91;])\n    ax.set_yticks(&#91;])    \n    if ind == 0:# \u8bbe\u7f6e\u603b\u6807\u9898\n        ax.set_title(titles&#91;3])\n    # \u8bbe\u7f6e\u6bcf\u4e2a\u67f1\u5f62\u56fe\u7684\u6570\u503c\u6807\u7b7e\u548cx\u8f74label\n    activity_values = &#91;'{:.1%}'.format(i) for i in each_data]\n    for i in range(len(x_ticket)):\n        ax.text(x_ticket&#91;i],each_data&#91;i]+0.05,s=activity_values&#91;i],ha='center',va='top')\n        ax.text(x_ticket&#91;i],-0.05,s=axtivity_labels&#91;i],ha='center',va='bottom')\n        \n# part 6  \u753b\u51fa\u6027\u522b\u5206\u5e03\nsex_data = features_all.iloc&#91;:,-3:]\nx_ticket = &#91;i for i in range(len(sex_data))]\nsex_labels = &#91;'SEX_{}'.format(i) for i in range(3)]\nylim_max = np.max(np.max(sex_data))\nfor ind,each_data in enumerate(sex_data.values):\n    ax = fig.add_subplot(line_index, col_index, ax_ids&#91;:,4]&#91;ind])\n    ax.bar(x=x_ticket,height=each_data) # \u753b\u67f1\u5f62\u56fe\n    ax.set_ylim((0, ylim_max*1.2))\n    ax.set_xticks(&#91;])\n    ax.set_yticks(&#91;])\n    if ind == 0: # \u8bbe\u7f6e\u6807\u9898\n       ax.set_title(titles&#91;4])    \n    # \u8bbe\u7f6e\u6bcf\u4e2a\u67f1\u5f62\u56fe\u7684\u6570\u503c\u6807\u7b7e\u548cx\u8f74label\n    sex_values = &#91;'{:.1%}'.format(i) for i in each_data]\n    for i in range(len(x_ticket)):\n        ax.text(x_ticket&#91;i],each_data&#91;i]+0.1,s=sex_values&#91;i],ha='center',va='top')\n        ax.text(x_ticket&#91;i],-0.1,s=sex_labels&#91;i],ha='center',va='bottom')\n    \nplt.tight_layout(pad=0.8) #\u8bbe\u7f6e\u9ed8\u8ba4\u7684\u95f4\u8ddd\n<\/code><\/pre>\n\n\n\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" loading=\"lazy\" width=\"1024\" height=\"720\" src=\"http:\/\/viplao.com\/wp-content\/uploads\/2024\/10\/image-25-1024x720.png\" alt=\"\" class=\"wp-image-2779\" srcset=\"http:\/\/viplao.com\/wp-content\/uploads\/2024\/10\/image-25-1024x720.png 1024w, http:\/\/viplao.com\/wp-content\/uploads\/2024\/10\/image-25-300x211.png 300w, http:\/\/viplao.com\/wp-content\/uploads\/2024\/10\/image-25-768x540.png 768w, http:\/\/viplao.com\/wp-content\/uploads\/2024\/10\/image-25-427x300.png 427w, http:\/\/viplao.com\/wp-content\/uploads\/2024\/10\/image-25.png 1473w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p><\/p>\n","protected":false},"excerpt":{"rendered":"<p># \u8bfb\u53d6\u6570\u636e raw_data = pd.read_csv(&#8216;cluster.txt&#038;#&hellip; <a href=\"http:\/\/viplao.com\/index.php\/2024\/10\/26\/%e6%95%b0%e6%8d%ae%e5%88%86%e6%9e%90-%e5%ae%a2%e6%88%b7%e7%89%b9%e5%be%81%e7%9a%84%e8%81%9a%e7%b1%bb%e5%88%86%e6%9e%90%e5%ae%9e%e8%b7%b5%e6%a1%88%e4%be%8b\/\" class=\"more-link read-more\" rel=\"bookmark\">\u7ee7\u7eed\u9605\u8bfb <span class=\"screen-reader-text\">\u6570\u5b57\u5316\u8fd0\u8425\u57fa\u7840\u6280\u80fd &#8211; \u5ba2\u6237\u7279\u5f81\u7684\u805a\u7c7b\u5206\u6790\u5b9e\u8df5\u6848\u4f8b<\/span><i class=\"fa fa-arrow-right\"><\/i><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[1],"tags":[28],"views":528,"_links":{"self":[{"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/posts\/2778"}],"collection":[{"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/comments?post=2778"}],"version-history":[{"count":2,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/posts\/2778\/revisions"}],"predecessor-version":[{"id":2804,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/posts\/2778\/revisions\/2804"}],"wp:attachment":[{"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/media?parent=2778"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/categories?post=2778"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/viplao.com\/index.php\/wp-json\/wp\/v2\/tags?post=2778"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}