{"id":644,"date":"2023-07-30T22:40:25","date_gmt":"2023-07-30T14:40:25","guid":{"rendered":"http:\/\/72.44.76.42\/?p=644"},"modified":"2023-07-30T22:40:25","modified_gmt":"2023-07-30T14:40:25","slug":"%e9%a2%84%e6%b5%8bboston%e6%88%bf%e4%bb%b7","status":"publish","type":"post","link":"https:\/\/www.notown.top\/?p=644","title":{"rendered":"\u9884\u6d4bBoston\u623f\u4ef7"},"content":{"rendered":"\n<h2 class=\"wp-block-heading\">-1.\u4f7f\u7528\u7684\u5e93<\/h2>\n\n\n\n<pre class=\"wp-block-code\"><code>import numpy as np\r\nimport pandas as pd\r\nfrom sklearn.model_selection import train_test_split\r\nfrom sklearn.preprocessing import StandardScaler\r\nfrom sklearn import linear_model\r\nfrom sklearn.metrics import mean_squared_error\r\nimport tensorflow as tf\r\nimport copy\n\ntf.random.set_seed(1234) # \u8bbe\u7f6e\u597d\u56fa\u5b9a\u7684\u968f\u673a\u6570\u79cd\u5b50\uff0c\u4ee5\u4fbf\u91cd\u590d\u5b9e\u9a8c<\/code><\/pre>\n\n\n\n<h2 class=\"wp-block-heading\">0.\u6570\u636e\u83b7\u53d6\u4e0e\u9884\u5904\u7406<\/h2>\n\n\n\n<pre class=\"wp-block-code\"><code># \u4e0b\u8f7d\u6570\u636e\u5e76\u6807\u51c6\u5316\r\ndef download_data():\r\n    data_url = \"http:\/\/lib.stat.cmu.edu\/datasets\/boston\"\r\n    raw_df = pd.read_csv(data_url, sep=\"\\s+\", skiprows=22, header=None)\r\n    data_x = np.hstack(&#91;raw_df.values&#91;::2, :], raw_df.values&#91;1::2, :2]])\r\n    data_y = raw_df.values&#91;1::2, 2]\r\n\r\n    data_x = StandardScaler().fit_transform(data_x)\r\n    # \u6570\u636e\u96c6\u5212\u5206\u6210\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6\r\n    train_x, test_x, train_y, test_y = train_test_split(data_x, data_y, test_size=0.1, random_state=1234)\r\n    # \u4ece\u8bad\u7ec3\u96c6\u4e2d\u5206\u51fa\u4e00\u70b9\u6765\u4f5c\u4e3a\u9a8c\u8bc1\u96c6\r\n    train_x, verify_x, train_y, verify_y = train_test_split(train_x, train_y, test_size=1\/9.0, random_state=1234)\r\n\r\n    return &#91;train_x, verify_x, test_x, train_y, verify_y, test_y] # \u8bad\u7ec3\u96c6\u3001\u9a8c\u8bc1\u96c6\u3001\u6d4b\u8bd5\u96c6\u3001\u8bad\u7ec3\u96c6\u7684y\u503c\u3001\u9a8c\u8bc1\u96c6\u7684y\u503c\u3001\u6d4b\u8bd5\u96c6\u7684y\u503c<\/code><\/pre>\n\n\n\n<h2 class=\"wp-block-heading\">1.\u7279\u5f81\u7684\u9009\u53d6<\/h2>\n\n\n\n<p>Boston\u623f\u4ef7\u6570\u636e\u670913\u4e2a\u7279\u5f81\uff0c\u53ea\u6709500\u591a\u4e2a\u6837\u4f8b\uff0c<strong>\u8fd9\u4e48\u591a\u7684\u7279\u5f81\u548c\u8fd9\u4e48\u5c11\u7684\u6570\u636e\u91cf<\/strong>\uff0c\u82e5\u4e0d\u5bf9\u7279\u5f81\u52a0\u4ee5\u7b5b\u9009\uff0c\u53ef\u80fd\u96be\u4ee5\u8bad\u7ec3\u51fa\u4e00\u4e2a\u4f18\u79c0\u7684\u6a21\u578b\u3002<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># \u8bbe\u7f6e\u7528\u4e8e\u8bad\u7ec3\u548c\u9884\u6d4b\u7684\u7279\u5f81\r\nprice_factor = &#91;True,          # \"CRIM\":\u72af\u7f6a\u7387\r\n                True,          # \"ZN\": \u8d85\u8fc725,000\u5e73\u65b9\u544e\u7684\u4f4f\u5b85\u7528\u5730\u6bd4\u4f8b\r\n                True,          # \"INDUS\":\u975e\u96f6\u552e\u5546\u4e1a\u9762\u79ef\u7684\u6bd4\u4f8b\r\n                True,           # \"CHAS\" \u67e5\u5c14\u65af\u6cb3\u9053\u9644\u8fd1\uff0c 1\u662f\uff0c0\u5426\r\n                True,          # \"NOX\":\u4e00\u6c27\u5316\u6c2e\u6d53\u5ea6(\u5343\u4e07\u5206\u4e4b\u4e00)\r\n                True,           # \"RM\": \u6bcf\u5957\u4f4f\u5b85\u7684\u5e73\u5747\u623f\u95f4\u6570\r\n                True,          # \"AGE\":1940\u5e74\u4ee5\u524d\u5efa\u9020\u7684\u81ea\u4f4f\u5355\u4f4d\u7684\u6bd4\u4f8b\r\n                True,          # \"DIS\": \u5230\u6ce2\u58eb\u987f\u4e94\u4e2a\u5c31\u4e1a\u4e2d\u5fc3\u7684\u52a0\u6743\u8ddd\u79bb\r\n                True,           # \"RAD\": \u5f84\u5411\u516c\u8def\u53ef\u8fbe\u6027\u6307\u6570\r\n                True,          # \"TAX\":\u6bcf1\u4e07\u7f8e\u5143\u7684\u5168\u989d\u623f\u4ea7\u7a0e\u7a0e\u7387\r\n                True,           # \"PTRATIO\":\u57ce\u9547\u5b66\u751f\u548c\u6559\u5e08\u7684\u6bd4\u4f8b\r\n                True,          # \"B\": \u5404\u57ce\u9547\u9ed1\u4eba\u7684\u6bd4\u4f8b\r\n                True           # \"LSTAT\": \u4eba\u53e3\u7684\u4f4e\u51fa\u751f\u7387\r\n                ]\n# \u5148\u5168\u9009\n\n# \u6309\u7167price_factor\u9009\u53d6\u7279\u5f81\r\ndef select_factors(data):\r\n    new_data = &#91;]\r\n    for i in range(3):\r\n        new_data.append(copy.deepcopy(data&#91;i]&#91;:, price_factor ])) # \u5bf9\u8bad\u7ec3\u96c6\u3001\u9a8c\u8bc1\u96c6\u3001\u6d4b\u8bd5\u96c6\u90fd\u53ea\u4fdd\u7559price_factor\u4e2d\u9009\u53d6\u7684\u7279\u5f81\r\n    for i in range(3, len(data)):\r\n        new_data.append(copy.deepcopy(data&#91;i]) ) # \u586b\u4e0a\u8bad\u7ec3\u96c6\u7684y\u503c\u3001\u9a8c\u8bc1\u96c6\u7684y\u503c\u3001\u6d4b\u8bd5\u96c6\u7684y\u503c\r\n    return new_data\n\n# \u4f7f\u7528sklearn\u5e93\u63d0\u4f9b\u7684\u5cad\u56de\u5f52\u8bad\u7ec3\u6a21\u578b\r\ndef linear_model_train(train_x, train_y):\r\n    my_model = linear_model.Ridge(alpha=0.06)\r\n    my_model.fit(train_x,train_y)\r\n    return my_model\r\n\nif __name__ == \"__main__\":\r\n    \r\n    data = download_data()\r\n    # find_best_feature(data)\r\n\r\n    train_x, verify_x, test_x ,train_y, verify_y, test_y = select_factors(data)\r\n    my_model = linear_model_train(train_x,train_y) # \u8bad\u7ec3\r\n    print(f\"train_data_perform: MSE \")\r\n    print(mean_squared_error(train_y, my_model.predict(train_x) ))\r\n    print(f\"verify_data_perform: MSE \")\r\n    print(mean_squared_error(verify_y, my_model.predict(verify_x) ))<\/code><\/pre>\n\n\n\n<p>\u76f4\u63a5\u5c06\u8fd913\u4e2a\u7279\u5f81\u90fd\u7528\u6765\u8bad\u7ec3\uff0c\u7ed3\u679c\u5982\u4e0b<br><code>train_data_perform: MSE<br>23.096444215174756<br>verify_data_perform: MSE<br>19.727919437983388<\/code><\/p>\n\n\n\n<p>\u5373\u8bad\u7ec3\u96c6\u7684\u5747\u65b9\u5dee\u4e3a23\uff0c\u9a8c\u8bc1\u96c6\u7684\u5747\u65b9\u5dee\u4e3a19\u3002\u4f5c\u4e3a\u4e00\u4e2a\u521d\u5b66\u8005\uff0c\u60f3\u5230\u7684\u6700\u7b80\u5355\u7c97\u66b4\u7684\u65b9\u6cd5\u5c31\u662f\uff0c<strong>\u5bf9\u6bcf\u79cd\u7279\u5f81\u7684\u7ec4\u5408\u90fd\u8bad\u7ec3\u4e00\u4e2a\u6a21\u578b\uff0c\u5e76\u4e14\u6bd4\u8f83\u5b83\u4eec\u9a8c\u8bc1\u96c6\u7684\u5747\u65b9\u5dee\u7ed3\u679c\uff0c<\/strong>\u5219\u5171\u4f1a\u67092^13-1\u79cd\u7ec4\u5408\uff0c\u51cf1\u662f\u56e0\u4e3a\u4e0d\u80fd\u4e00\u4e2a\u7279\u5f81\u90fd\u4e0d\u9009\u3002\u8fd9\u4e2a\u53ef\u4ee5\u7528\u9012\u5f52\u56de\u6eaf\u6cd5\u5b9e\u73b0\u3002<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">2. \u6a21\u578b\u7684\u6bd4\u8f83<\/h2>\n\n\n\n<p>\u8fd9\u4e2a\u6a21\u578b\u7684\u529f\u80fd\u5f88\u7b80\u5355\uff0c\u5c31\u662f\u9884\u6d4b\u623f\u4ef7\uff0c\u6700\u7b80\u5355\u7684\u5c31\u662f\u76f4\u63a5\u4f7f\u7528sklearn\u5305\u7684\u4e00\u4e2a\u7ebf\u6027\u56de\u5f52\u5355\u5143\u5b8c\u6210\uff0c\u4f46\u6211\u8fd8\u60f3\u7528\u4e00\u4e2a\u5c0f\u7684\u795e\u7ecf\u7f51\u7edc\u8fdb\u884c\u6bd4\u8f83\u4e00\u4e0b\u6548\u679c\uff0c\u4e8e\u662f\u8fd9\u91cc\u6211\u4f7f\u7528\u4e86sklearn\u5305\u91cc\u7ebf\u6027\u56de\u5f52\u4e2d\u7684\u5cad\u56de\u5f52\uff0c\u5e76\u4e14\u4f7f\u7528tensorflow\u4e2d\u7684Sequential\u521b\u5efa\u4e00\u4e2a\u6a21\u578b\u3002\u5cad\u56de\u5f52\u7684\u4ee3\u7801\u5982\u4e0a\uff0c\u795e\u7ecf\u7f51\u7edc\u7684\u4ee3\u7801\u5982\u4e0b\uff1a<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># \u521b\u5efa\u4e00\u4e2a\u5c0f\u795e\u7ecf\u7f51\u7edc\r\ndef creat_deep_learn_model():\r\n    # \u521b\u5efa\u6a21\u578b\u65f6\u6211\u76f4\u63a5\u6307\u5b9a\u4e86\u53c2\u6570\u7684\u521d\u503c\uff0c\u82e5\u4e0d\u6307\u5b9a\uff0c\u5219\u4f1a\u968f\u673a\u4ea7\u751f\uff0c\u5bf9\u4e8e\u540e\u7eed\u6311\u9009\u8868\u73b0\u6700\u597d\u7684\u7279\u5f81\u9020\u6210\u5f71\u54cd\uff08\u5373\u4fdd\u6301\u5355\u4e00\u53d8\u91cf\uff0c\u53ea\u80fd\u53d8\u7279\u5f81\uff09\r\n    my_model = tf.keras.models.Sequential(&#91;\r\n        tf.keras.layers.Dense(units=5,activation=\"relu\",kernel_initializer=tf.keras.initializers.Ones(),bias_initializer=tf.keras.initializers.Ones() ),\r\n        tf.keras.layers.Dense(units=1,activation=\"relu\",kernel_initializer=tf.keras.initializers.Ones(),bias_initializer=tf.keras.initializers.Ones() )\r\n    ])\r\n    my_model.compile(\r\n        loss=\"mse\",\r\n        optimizer=tf.keras.optimizers.Adam(0.06)\r\n    )\r\n    return my_model\r\n\r\n# \u8bad\u7ec3\u795e\u7ecf\u7f51\u7edc\r\ndef deep_learning_train(train_x, train_y):\r\n    my_model = creat_deep_learn_model()\r\n    my_model.fit(train_x, train_y, epochs=40, use_multiprocessing=True, workers=4, batch_size=128, verbose=0)\r\n    return my_model<\/code><\/pre>\n\n\n\n<p>\u8fd9\u91cc\u8c03\u7528fit\u51fd\u6570\u65f6\uff0c\u5c06batch_size\u8c03\u6574\u5230128\uff0c\u4ee5\u52a0\u5feb\u6211\u4eec\u9009\u53d6\u8bad\u7ec3\u901f\u5ea6\uff08\u540e\u9762\u6311\u9009\u7279\u5f81\u8981\u8bad\u7ec3\u516b\u5343\u591a\u4e2a\u6a21\u578b\uff09<\/p>\n\n\n\n<p>\u521b\u5efa\u4e00\u4e2a\u51fd\u6570\u6765\u68c0\u9a8c\u4e24\u79cd\u6a21\u578b\u7684\u8868\u73b0<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code># \u68c0\u67e5\u5cad\u56de\u5f52\u548c\u795e\u7ecf\u7f51\u7edc\u5728\u8bad\u7ec3\u96c6\u548c\u9a8c\u8bc1\u96c6\u7684\u8868\u73b0\r\ndef check_perform(model_list, train_x, train_y, verify_x, verify_y):\r\n    print(f\"train_data_perform: MSE \")\r\n    for i in range(len(model_list) ):\r\n        print(f\"{i} : {mean_squared_error(train_y, model_list&#91;i].predict(train_x) )}\")\r\n\r\n    print(f\"verify_data_perform: MSE \")\r\n    for i in range(len(model_list) ):\r\n        pred_y = model_list&#91;i].predict(verify_x)\r\n        print(f\"{i} : {mean_squared_error(verify_y, pred_y)}\")<\/code><\/pre>\n\n\n\n<p>main\u91cc\u6539\u4e00\u6539<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>if __name__ == \"__main__\":\r\n    \r\n    data = download_data()\r\n    # find_best_feature(data)\r\n\r\n    train_x, verify_x, test_x ,train_y, verify_y, test_y = select_factors(data)\r\n    model_list = &#91;]\r\n    model_list.append(linear_model_train(train_x,train_y))\r\n\r\n    model_list.append(deep_learning_train(train_x,train_y))\r\n\r\n    check_perform(model_list, train_x, train_y, verify_x, verify_y)<\/code><\/pre>\n\n\n\n<p>\u4e24\u4e2a\u6a21\u578b\u8868\u73b0\u5982\u4e0b\uff1a<br><code>train_data_perform: MSE<br>0 : 23.096444215174756<br>13\/13 [==============================] - 0s 2ms\/step<br>1 : 30.51652489850927<br>verify_data_perform: MSE<br>0 : 19.727919437983388<br>2\/2 [==============================] - 0s 1ms\/step<br>1 : 17.252091274575182<\/code><\/p>\n\n\n\n<p>\u521d\u6b65\u89c2\u5bdf\uff0c\u5728\u8bad\u7ec3\u96c6\u91cc\uff0c\u4e00\u4e2a\u7b80\u5355\u7684\u7ebf\u6027\u56de\u5f52\u8981\u6bd4\u795e\u7ecf\u7f51\u7edc\u8868\u73b0\u8981\u597d\uff0c\u4f46\u5728\u9a8c\u8bc1\u96c6\u91cc\uff0c\u53c8\u8981\u5dee\u70b9\uff0c\u73b0\u5728\u53ef\u80fd\u8fd8\u96be\u4ee5\u5206\u6e05\u8c01\u597d\u8c01\u574f\u3002\u6211\u4eec\u5f00\u59cb\u9009\u53d6\u6211\u4eec\u7684\u6700\u4f73\u7279\u5f81<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">3.\u4f7f\u7528\u9012\u5f52\u56de\u6eaf\u7b97\u51fa2^13-1\u79cd\u7279\u5f81\u7ec4\u5408\u91cc\uff0c\u6700\u4f4e\u7684\u5747\u65b9\u5dee<\/h2>\n\n\n\n<pre class=\"wp-block-code\"><code># \u8ba1\u7b97\u5cad\u56de\u5f52\u548c\u795e\u7ecf\u7f51\u7edc\u7684\u9a8c\u8bc1\u96c6\u7684\u5747\u65b9\u5dee\r\ndef get_mse(data):\r\n    train_x, verify_x, test_x ,train_y, verify_y, test_y = select_factors(data)\r\n    a = linear_model_train(train_x,train_y)\r\n    b = deep_learning_train(train_x,train_y)\r\n    return mean_squared_error(verify_y, a.predict(verify_x) ), mean_squared_error(verify_y, b(verify_x) )\r\n\r\nmin_linear_err = 10000000.0             # \u5cad\u56de\u5f52\u6700\u5c0f\u5747\u65b9\u5dee\u521d\u59cb\u503c\r\nmin_deep_learn_err = 10000000.0         # \u795e\u7ecf\u7f51\u7edc\u6700\u5c0f\u5747\u65b9\u5dee\u521d\u59cb\u503c\r\nmin_linear_features = &#91;]                # \u5cad\u56de\u5f52\u6700\u5c0f\u5747\u65b9\u5dee\u65f6\u7684\u7279\u5f81\u9009\u53d6\r\nmin_deep_learn_features = &#91;]            # \u795e\u7ecf\u7f51\u7edc\u6700\u5c0f\u5747\u65b9\u5dee\u65f6\u7684\u7279\u5f81\u9009\u53d6\r\nall_zero = True                         # all_zero\u4e3atrue\u65f6\u8868\u793a\u4e00\u4e2a\u7279\u5f81\u4e5f\u4e0d\u9009\r\n\r\n# \u56de\u6eaf\u6cd5\u627e\u6bcf\u4e00\u79cd\u7279\u5f81\u9009\u53d6\u7684\u53ef\u80fd\u4ece\u5168False\u5230\u5168True\uff0c\u5171\u67092^13\u79cd\u7ec4\u5408\r\ndef find_func(idx, data):\r\n    global min_linear_err, min_deep_learn_err, min_linear_features, min_deep_learn_features, all_zero\r\n    length = len( price_factor )\r\n    if idx >= length:                   # \u586b\u5145\u5b8c\u6574\u4e2aprice_factor\u540e\r\n        if all_zero:                    # \u9996\u6b21\u8fdb\u6765\u80af\u5b9a\u5168False\uff0c\u8981return\u6389\uff0c\u603b\u4e0d\u80fd\u4e00\u4e2a\u7279\u5f81\u90fd\u6ca1\u6709\r\n            all_zero = False            # \u540e\u9762\u5c31\u6539\u6210Fales\r\n            return\r\n        # \u4e0d\u662f\u7b2c\u4e00\u6b21\u8fdb\u6765\u5c31\u80af\u5b9a\u975e\u5168False\r\n        a,b = get_mse(data)     # \u83b7\u53d6\u5b83\u4eec\u7684\u7ebf\u6027\u56de\u5f52\u548c\u795e\u7ecf\u7f51\u7edc\u6a21\u578b\u7684\u5747\u65b9\u8bef\u5dee\r\n        if a &lt; min_linear_err:\r\n            min_linear_err = a\r\n            min_linear_features = copy.deepcopy(price_factor)\r\n            print(f\"linear mse = {a}, {price_factor}\")\r\n        if b &lt; min_deep_learn_err:\r\n            min_deep_learn_err = b\r\n            min_deep_learn_features = copy.deepcopy(price_factor)\r\n            print(f\"deep_learn mse = {b}, {price_factor}\")\r\n        return\r\n    price_factor&#91;idx] = False\r\n    find_func(idx+1,data)\r\n    price_factor&#91;idx] = True\r\n    find_func(idx+1,data)\r\n\r\n# \u627e\u6700\u4f73\u7684\u7279\u5f81\u9009\u53d6\r\ndef find_best_feature(data):\r\n    for i in range(len(price_factor)):\r\n        price_factor&#91;i] = False\r\n    find_func(0, data)\r\n    print(f\"linear_min_mse = {min_linear_err}, features= {min_linear_features}\")\r\n    print(f\"deepLearn_min_mse = {min_deep_learn_err}, features= {min_deep_learn_features}\")<\/code><\/pre>\n\n\n\n<p>main\u91cc\u9762\u6539\u6210\u8fd9\u6837\u5b50<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>if __name__ == \"__main__\":\r\n    \r\n    data = download_data()\r\n    find_best_feature(data)<\/code><\/pre>\n\n\n\n<p>\u6700\u540e\u8dd1\u4e86\u4e2a\u628a\u5c0f\u65f6\uff0c\u8f93\u51fa\u7684\u9a8c\u8bc1\u96c6\u7684\u6700\u5c0f\u5747\u65b9\u5dee\u4e3a\uff1a<br><code>linear_min_mse = 14.759951009440611, features= [True, False, False, True, False, True, True, True, True, True, True, False, False]<br>deepLearn_min_mse = 11.417357419353857, features= [True, False, True, False, True, True, True, True, True, True, True, True, False]<\/code><\/p>\n\n\n\n<p>\u53ef\u89c1\u901a\u8fc7<strong>\u6311\u9009\u51fa\u90e8\u5206\u76f8\u5173\u6027\u66f4\u5f3a\u7279\u5f81\u7ec4\u5408\u5728\u4e00\u8d77\uff0c\u6700\u540e\u7684\u6548\u679c\u4f1a\u66f4\u597d\uff0c\u7279\u522b\u662f\u6570\u636e\u5c11\u7684\u60c5\u51b5<\/strong>\u4e0b\u3002\u6211\u4eec\u518d\u6309\u795e\u7ecf\u7f51\u7edc\u6a21\u578b\u6700\u5c0f\u5747\u65b9\u5dee\u65f6\u7684\u9009\u51fa\u7684\u7279\u5f81\uff0c\u5c1d\u8bd5\u4fee\u6539\u4e00\u4e0b\u6a21\u578b\u53c2\u6570\u8fdb\u4e00\u6b65\u8bad\u7ec3\u3002<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">4. \u8c03\u6574\u53c2\u6570\u91cd\u65b0\u8bad\u7ec3\u795e\u7ecf\u7f51\u7edc<\/h2>\n\n\n\n<p>\u5c06\u7f16\u8bd1\u6a21\u578b\u65f6\u4f7f\u7528\u7684<strong>\u5b66\u4e60\u7387\u521d\u503c\u8bbe\u4f4e\u70b9<\/strong>\uff0c\u59820.04\uff0c\u5c06\u8bad\u7ec3\u6a21\u578b\u7684\u7684<strong>\u8f6e\u6b21\u589e\u52a0\u52301000\u6b21<\/strong>\uff0c<strong>batch_size\u8c03\u523032<\/strong><\/p>\n\n\n\n<p>\u53ef\u5f97\u5230\uff0c\u5f53\u7279\u5f81\u9009\u53d6\u4e3a\uff1a<br><code>True, False, True, False, True, True, True, True, True, True, True, True, False<\/code><\/p>\n\n\n\n<p>\u795e\u7ecf\u7f51\u7edc\u8bad\u7ec3\u96c6\u7684\u5747\u65b9\u5dee\u4e3a\uff1a9.53064629681389\uff0c\u9a8c\u8bc1\u96c6\u7684\u5747\u65b9\u5dee\u4e3a7.670203426797456<br>\uff08\u6bcf\u6b21\u8bad\u7ec3\u53ef\u80fd\u5f97\u5230\u7684\u7ed3\u679c\u4f1a\u6d6e\u52a8\uff0c\u4f46\u5dee\u8ddd\u4e0d\u4f1a\u592a\u5927\uff09<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">5.\u5b9e\u9a8c\u4e2d\u5b58\u5728\u7684\u95ee\u9898<\/h2>\n\n\n\n<p>\u800c\u5f53\u7279\u5f81\u9009\u53d6\u4e3a\uff1a<br><code>True, False, False, True, False, True, True, True, True, True, True, False, False<\/code><\/p>\n\n\n\n<p>\u5cad\u56de\u5f52\u7684\u8bad\u7ec3\u96c6\u5747\u65b9\u5dee\u4e3a\uff1a34.23614743088512\uff0c\u9a8c\u8bc1\u96c6\u5747\u65b9\u5dee\u4e3a\uff1a14.759951009440611<\/p>\n\n\n\n<p>\u53ef\u89c1\u8fd9\u4e2a\u5cad\u56de\u5f52\u91cc\u8bad\u7ec3\u96c6\u8868\u73b0\u4e0d\u597d\uff0c\u9a8c\u8bc1\u96c6\u8868\u73b0\u597d\uff0c\u4f46\u56e0\u4e3a\u8fd9\u4e2a\u9009\u6700\u4f18\u7279\u5f81\u7ec4\u5408\u7684\uff0c\u5c31\u662f\u6309\u9a8c\u8bc1\u96c6\u5747\u65b9\u5dee\u6700\u5c11\u7684\u6765\u9009\uff0c\u6545\u4e0d\u592a\u51c6\u786e\u3002\u4f46\u672c\u6b21\u5b9e\u9a8c\u4e2d\uff0c\u6211\u4eec\u8fd8\u662f\u53ef\u4ee5\u5f97\u5230\uff0c<strong>\u7ecf\u8fc7\u8c03\u6574\u540e\u7684\u795e\u7ecf\u7f51\u7edc\u8fd8\u662f\u8981\u6bd4\u5355\u4e00\u7684\u7ebf\u6027\u56de\u5f52\u5355\u5143\u6765\u9884\u6d4b\u623f\u4ef7\uff0c\u8981\u66f4\u52a0\u51c6\u786e\u7684<\/strong>\u3002<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">6.\u67e5\u770b\u6a21\u578b\u5728\u6d4b\u8bd5\u96c6\u4e0a\u7684\u7ed3\u679c<\/h2>\n\n\n\n<p>\u5728\u4e3a\u5cad\u56de\u5f52\u9009\u53d6\u6700\u4f73\u7279\u5f81\u7ec4\u5408\u540e\uff0c\u5176\u6d4b\u8bd5\u96c6\u4e0a\u7684\u5747\u65b9\u5dee\u4e3a\uff1a<br><code>test linear mse :14.14316351410805<\/code>\uff0c\u4e0e\u5cad\u56de\u5f52\u9a8c\u8bc1\u96c6\u5747\u65b9\u5dee\u5dee\u4e0d\u591a<\/p>\n\n\n\n<p>\u5728\u4e3a\u795e\u7ecf\u7f51\u7edc\u9009\u53d6\u6700\u4f73\u7279\u5f81\u7ec4\u5408\u540e\uff0c\u5176\u5728\u6d4b\u8bd5\u96c6\u4e0a\u7684\u5747\u65b9\u5dee\u4e3a\uff1a<br><code>test deepLearn mse :10.750304093169447<\/code>\uff0c\u6bd4\u795e\u7ecf\u7f51\u7edc\u7684\u9a8c\u8bc1\u96c6\u7684\u5747\u65b9\u5dee\u7565\u4f4e<\/p>\n\n\n\n<p>\u53ef\u89c1\u6700\u540e\u7684\u7ed3\u679c\u4f7f\u7528\u4e00\u4e2a\u5c0f\u89c4\u6a21\u7684\u795e\u7ecf\u7f51\u7edc\u8fd8\u662f\u8981\u6bd4\u7ebf\u6027\u56de\u5f52\u5355\u5143\u8981\u8868\u73b0\u66f4\u4f18\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>-1.\u4f7f\u7528\u7684\u5e93 0.\u6570\u636e\u83b7\u53d6\u4e0e\u9884\u5904\u7406 1.\u7279\u5f81\u7684\u9009\u53d6 Boston\u623f\u4ef7\u6570\u636e\u670913\u4e2a\u7279\u5f81\uff0c\u53ea\u6709500\u591a\u4e2a\u6837\u4f8b\uff0c\u8fd9&hellip;<a href=\"https:\/\/www.notown.top\/?p=644\" class=\"more-link\">Continue reading <span class=\"screen-reader-text\">\u9884\u6d4bBoston\u623f\u4ef7<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_import_markdown_pro_load_document_selector":0,"_import_markdown_pro_submit_text_textarea":"","footnotes":""},"categories":[19],"tags":[],"class_list":["post-644","post","type-post","status-publish","format-standard","hentry","category-19"],"_links":{"self":[{"href":"https:\/\/www.notown.top\/index.php?rest_route=\/wp\/v2\/posts\/644","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.notown.top\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.notown.top\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.notown.top\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.notown.top\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=644"}],"version-history":[{"count":6,"href":"https:\/\/www.notown.top\/index.php?rest_route=\/wp\/v2\/posts\/644\/revisions"}],"predecessor-version":[{"id":650,"href":"https:\/\/www.notown.top\/index.php?rest_route=\/wp\/v2\/posts\/644\/revisions\/650"}],"wp:attachment":[{"href":"https:\/\/www.notown.top\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=644"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.notown.top\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=644"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.notown.top\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=644"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}