{"id":122,"date":"2024-09-24T21:08:41","date_gmt":"2024-09-24T13:08:41","guid":{"rendered":"http:\/\/liutx.xyz\/?p=122"},"modified":"2024-09-24T21:13:40","modified_gmt":"2024-09-24T13:13:40","slug":"%e7%89%b9%e5%be%81%e9%80%89%e6%8b%a9%e4%b8%8e%e6%98%be%e7%a4%ba%e4%bb%a3%e7%a0%81","status":"publish","type":"post","link":"http:\/\/liutx.xyz\/?p=122","title":{"rendered":"\u7279\u5f81\u9009\u62e9\u4e0e\u663e\u793a\u4ee3\u7801"},"content":{"rendered":"\n<pre class=\"wp-block-code\"><code>\nimport pandas as pd\nimport matplotlib.pyplot as plt\nimport random\nfrom sklearn import tree\nfrom sklearn.model_selection import cross_val_score\n\ndata = pd.read_csv('.\/dataset\/sonar.all-data',header=None,sep=',')\nprint(data.head())\nX = data.iloc&#91;:,:-1]\ny = data.iloc&#91;:,-1:].values.flatten()\n\niterations = 100 # \u8fed\u4ee3\u6b21\u6570\npop_size = 100   # \u79cd\u7fa4\u5927\u5c0f\uff0c\u591a\u5c11\u4e2a\u67d3\u8272\u4f53\npc = 0.25   # \u4ea4\u53c9\u6982\u7387\npm = 0.01   # \u53d8\u5f02\u6982\u7387\n\nchrom_length = len(data.columns)-1    # \u67d3\u8272\u4f53\u957f\u5ea6\npop = &#91;]    # \u79cd\u7fa4\nfitness_list = &#91;]   # \u9002\u5e94\u5ea6\nratio_list = &#91;]     # \u7d2f\u8ba1\u6982\u7387\n\n\n# \u521d\u59cb\u5316\u79cd\u7fa4\ndef geneEncoding():\n    i = 0\n    while i &lt; pop_size:\n        temp = &#91;]\n        has_1 = False   # \u8fd9\u6761\u67d3\u8272\u4f53\u662f\u5426\u67091\n        for j in range(chrom_length):\n            rand = random.randint(0,1)\n            if rand == 1:\n                has_1 = True\n            temp.append(rand)\n        if has_1:   # \u67d3\u8272\u4f53\u4e0d\u80fd\u51680\n            i += 1\n            pop.append(temp)\n        \n\n# \u8ba1\u7b97\u9002\u5e94\u5ea6\ndef calFitness():\n    fitness_list.clear()\n    for i in range(pop_size):   # \u8ba1\u7b97\u79cd\u7fa4\u4e2d\u6bcf\u6761\u67d3\u8272\u4f53\u7684\u9002\u5e94\u5ea6\n        X_test = X\n\n        has_1 = False\n        for j in range(chrom_length):\n            if pop&#91;i]&#91;j] == 0:\n                X_test =X_test.drop(columns = j)\n            else:\n                has_1 = True\n        X_test = X_test.values\n        \n        if has_1:\n            clf = tree.DecisionTreeClassifier() # \u51b3\u7b56\u6811\u4f5c\u4e3a\u5206\u7c7b\u5668\n            fitness = cross_val_score(clf, X_test, y, cv=5).mean()  # 5\u6b21\u4ea4\u53c9\u9a8c\u8bc1\n            fitness_list.append(fitness)\n        else:\n            fitness = 0     # \u51680\u7684\u9002\u5e94\u5ea6\u4e3a0\n            fitness_list.append(fitness)\n\n# \u8ba1\u7b97\u9002\u5e94\u5ea6\u7684\u603b\u548c\ndef sumFitness():\n    total = 0\n    for i in range(pop_size):\n        total += fitness_list&#91;i]\n    return total\n\n# \u8ba1\u7b97\u6bcf\u6761\u67d3\u8272\u4f53\u7684\u7d2f\u8ba1\u6982\u7387\ndef getRatio():\n    ratio_list.clear()\n    ratio_list.append(fitness_list&#91;0])\n    for i in range(1, pop_size):\n        ratio_list.append(ratio_list&#91;i-1] + fitness_list&#91;i])\n    ratio_list&#91;-1] = 1\n\n# \u9009\u62e9\ndef selection():\n    global pop\n    total_fitness = sumFitness()\n    for i in range(pop_size):\n        fitness_list&#91;i] = fitness_list&#91;i] \/ total_fitness\n    getRatio()\n    \n    rand_ratio = &#91;] # \u968f\u673a\u6982\u7387\n    for i in range(pop_size):\n        rand_ratio.append(random.random())\n    rand_ratio.sort()\n\n    new_pop = &#91;]    # \u65b0\u79cd\u7fa4\n    i = 0  # \u5df2\u7ecf\u5904\u7406\u7684\u968f\u673a\u6982\u7387\u6570\n    j = 0  # \u8d85\u51fa\u8303\u56f4\u7684\u67d3\u8272\u4f53\u6570\n   \n    while i &lt; pop_size:\n        if rand_ratio&#91;i] &lt; ratio_list&#91;j]:   # \u968f\u673a\u6570\u5728\u7b2cj\u4e2a\u67d3\u8272\u4f53\u7684\u6982\u7387\u8303\u56f4\u5185\n            new_pop.append(pop&#91;j])\n            i += 1\n        else:\n            j += 1\n\n    pop = new_pop\n\n# \u4ea4\u53c9\ndef crossover():\n    for i in range(pop_size-1): # \u82e5\u4ea4\u53c9\uff0c\u5219\u67d3\u8272\u4f53i\u4e0e\u67d3\u8272\u4f53i+1\u4ea4\u53c9\n        if random.random() &lt; pc:# \u53d1\u751f\u4ea4\u53c9\n            cpoint = random.randint(0, chrom_length-1)    # \u968f\u673a\u9009\u62e9\u4ea4\u53c9\u70b9\n            temp1 = &#91;]\n            temp2 = &#91;]\n            temp1.extend(pop&#91;i]&#91;:cpoint])\n            temp1.extend(pop&#91;i+1]&#91;cpoint:])\n            temp2.extend(pop&#91;i+1]&#91;:cpoint])\n            temp2.extend(pop&#91;i]&#91;cpoint:])\n            pop&#91;i] = temp1\n            pop&#91;i+1] = temp2\n\n# \u53d8\u5f02\ndef mutation():\n    for i in range(pop_size):\n        if random.random() &lt; pm: # \u53d1\u751f\u53d8\u5f02\n            mpoint = random.randint(0, chrom_length-1)  # \u968f\u673a\u9009\u62e9\u53d8\u5f02\u70b9\n            if pop&#91;i]&#91;mpoint] == 1:\n                pop&#91;i]&#91;mpoint] = 0\n            else:\n                pop&#91;i]&#91;mpoint] = 1\n\n# \u6700\u4f18\u89e3\ndef getBest():\n    best_chrom = pop&#91;0]\n    best_fitness = fitness_list&#91;0]\n    for i in range(1,pop_size):\n        if fitness_list&#91;i] &gt; best_fitness:\n            best_fitness = fitness_list&#91;i]  # \u6700\u4f73\u9002\u5e94\u503c\n            best_chrom = pop&#91;i] # \u6700\u4f73\u67d3\u8272\u4f53\n\n    return best_chrom, best_fitness\n\nif __name__=='__main__':\n\n    px = &#91;]\n    py = &#91;]\n    plt.ion()\n\n    results = &#91;]\n    geneEncoding() # \u521d\u59cb\u5316\u79cd\u7fa4\n    for i in range(iterations):\n        print(i)\n\n        calFitness() # \u8ba1\u7b97\u79cd\u7fa4\u4e2d\u6bcf\u6761\u67d3\u8272\u4f53\u9002\u5e94\u5ea6\n\n        best_chrom, best_fitness = getBest()\n        results.append(&#91;i, best_chrom, best_fitness])\n\n        selection() # \u9009\u62e9\n        crossover() # \u4ea4\u53c9\n        mutation()  # \u53d8\u5f02\n        plt.title('GA feature selection')\n        plt.xlabel('iterations')\n        plt.ylabel('best fitness')\n        plt.xlim((0, iterations))  # x\u5750\u6807\u8303\u56f4\n        plt.ylim((0.6, 0.8))  # y\u5750\u6807\u8303\u56f4\n\n        print(&#91;i, best_chrom, best_fitness])\n\n        px.append(i)    # \u753b\u56fe\n        py.append(best_fitness)\n        plt.plot(px,py)\n        plt.show()\n        plt.pause(0.001)<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">\u5229\u7528\u9057\u4f20\u7b97\u6cd5\u505a\u7279\u5f81\u9009\u62e9\uff0c\u5e76\u7ed8\u56fe\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e0b\u9762\u662f\u5229\u7528t-SNE\u7b97\u6cd5\u8fdb\u884c\u7279\u5f81\u964d\u7ef4\u5e76\u5c55\u793a\uff1a<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>import pandas as pd\nfrom sklearn.manifold import TSNE\nimport matplotlib.pyplot as plt\nfrom mpl_toolkits.mplot3d import Axes3D\n\n# \u52a0\u8f7d\u6570\u636e\u96c6\uff08\u8fd9\u91cc\u4ee5sonar.all-data\u4e3a\u4f8b\uff0c\u4f46\u8bf7\u786e\u4fdd\u8def\u5f84\u6b63\u786e\uff09\ndata = pd.read_csv('.\/dataset\/sonar.all-data', header=None, sep=',')\n# \u7ed9\u5b9a\u7684\u6570\u7ec4\uff0c\u8fd9\u662f\u9057\u4f20\u7b97\u6cd5\u505a\u7279\u5f81\u9009\u62e9\u7684\u7ed3\u679c\narr = &#91;0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0,1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 1, 1, 1, 1, 1, 0, 1]\n# \u4f7f\u7528\u5217\u8868\u63a8\u5bfc\u5f0f\u627e\u51fa\u6240\u6709\u6807\u7b7e\u4e3a1\u7684\u7d22\u5f15\nindices_of_ones = &#91;i for i, value in enumerate(arr) if value == 1]\n# \u8f93\u51fa\u7ed3\u679c\nprint(indices_of_ones)\n# \u5047\u8bbe\u6700\u540e\u4e00\u5217\u662f\u6807\u7b7e\uff0c\u5176\u4f59\u662f\u7279\u5f81\nX = data.iloc&#91;:, indices_of_ones]  # \u7279\u5f81\u6570\u636e\ny = data.iloc&#91;:, -1]  # \u6807\u7b7e\u6570\u636e\uff08\u6ce8\u610f\u8fd9\u91cc\u6211\u4eec\u4e0d\u9700\u8981flatten\uff0c\u56e0\u4e3ay\u5df2\u7ecf\u662fSeries\uff09\n\n# \u521d\u59cb\u5316t-SNE\u6a21\u578b\n# \u4f60\u53ef\u4ee5\u8c03\u6574perplexity\u548cn_iter\u7b49\u53c2\u6570\u4ee5\u4f18\u5316\u7ed3\u679c\n# \u6ce8\u610f\uff1a\u7531\u4e8en_iter\u5728scikit-learn\u7684\u8f83\u65b0\u7248\u672c\u4e2d\u5df2\u88ab\u5f03\u7528\u5e76\u66ff\u6362\u4e3amax_iter\uff0c\u8fd9\u91cc\u6211\u4eec\u4f7f\u7528max_iter\ntsne = TSNE(n_components=3, perplexity=30, max_iter=300, random_state=42)\n\n# \u6267\u884ct-SNE\u964d\u7ef4\nX_tsne = tsne.fit_transform(X)\n\n# \u53ef\u89c6\u5316\u7ed3\u679c\n# \u8fd9\u91cc\u6211\u4eec\u4f7f\u7528\u6563\u70b9\u56fe\u6765\u53ef\u89c6\u5316\u4e0d\u540c\u7c7b\u522b\u7684\u6570\u636e\u70b9\nplt.figure(figsize=(8, 6))\nlabel_to_color = {'R': 'red', 'M': 'blue'}\nfor label in y.unique():\n    # \u7b5b\u9009\u51fa\u5c5e\u4e8e\u5f53\u524d\u7c7b\u522b\u7684\u6570\u636e\u70b9\n    points = X_tsne&#91;y == label]\n    # \u7ed8\u5236\u6563\u70b9\u56fe\n    plt.scatter(points&#91;:, 0], points&#91;:, 1], color=label_to_color&#91;label], label=label)\n\nplt.legend()\nplt.title('t-SNE Visualization of Sonar Dataset')\nplt.xlabel('t-SNE Feature 1')\nplt.ylabel('t-SNE Feature 2')\nplt.show()\n<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e0b\u9762\u662f\u7ed8\u5236\u6df7\u6dc6\u77e9\u9635\uff1a<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>\nfrom sklearn.metrics import confusion_matrix\nimport matplotlib.pyplot as plt\nimport seaborn as sns\n\n# # \u5047\u8bbe\u8fd9\u662f\u4f60\u7684\u6a21\u578b\u9884\u6d4b\u7ed3\u679c\n# y_pred = &#91;2, 0, 2, 2, 0, 1, 1, 2]\n# # \u5047\u8bbe\u8fd9\u662f\u4f60\u7684\u771f\u5b9e\u7ed3\u679c\n# y_true = &#91;2, 0, 2, 2, 0, 1, 0, 2]\n\n# \u8ba1\u7b97\u6df7\u6dc6\u77e9\u9635\ncm = confusion_matrix(y_true, y_pred)\nprint(cm)\n\n# \u7ed8\u5236\u6df7\u6dc6\u77e9\u9635\nclasses = &#91;'other', 'frogman', 'UUV']\nsns.heatmap(cm, annot=True, fmt=\"d\", cmap=\"Blues\", xticklabels=classes, yticklabels=classes, annot_kws={\"size\": 16})\nplt.ylabel('True label')\nplt.xlabel('Predicted label')\nplt.title('Confusion Matrix')\n# plt.title('\u8f68\u8ff9\u7279\u5f81\u6df7\u6dc6\u77e9\u9635')\nplt.show()<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"<p>\u5229\u7528\u9057\u4f20\u7b97\u6cd5\u505a\u7279\u5f81\u9009\u62e9\uff0c\u5e76\u7ed8\u56fe\u3002 \u4e0b\u9762\u662f\u5229\u7528t-SNE\u7b97\u6cd5\u8fdb\u884c\u7279\u5f81\u964d\u7ef4\u5e76\u5c55\u793a\uff1a \u4e0b\u9762\u662f\u7ed8\u5236\u6df7\u6dc6\u77e9\u9635\uff1a<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-122","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"http:\/\/liutx.xyz\/index.php?rest_route=\/wp\/v2\/posts\/122","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/liutx.xyz\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/liutx.xyz\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/liutx.xyz\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/liutx.xyz\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=122"}],"version-history":[{"count":3,"href":"http:\/\/liutx.xyz\/index.php?rest_route=\/wp\/v2\/posts\/122\/revisions"}],"predecessor-version":[{"id":125,"href":"http:\/\/liutx.xyz\/index.php?rest_route=\/wp\/v2\/posts\/122\/revisions\/125"}],"wp:attachment":[{"href":"http:\/\/liutx.xyz\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=122"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/liutx.xyz\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=122"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/liutx.xyz\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=122"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}