{"id":9470,"date":"2025-11-07T14:00:52","date_gmt":"2025-11-07T06:00:52","guid":{"rendered":"\/?p=9470"},"modified":"2025-11-07T14:00:52","modified_gmt":"2025-11-07T06:00:52","slug":"hanlp-%e5%9f%ba%e4%ba%8esvm%e6%94%af%e6%8c%81%e5%90%91%e9%87%8f%e6%9c%ba-%e5%ae%9e%e8%b7%b5-%e6%96%87%e6%9c%ac%e5%88%86%e7%b1%bb","status":"publish","type":"post","link":"\/?p=9470","title":{"rendered":"HanLP \u57fa\u4e8eSVM\u652f\u6301\u5411\u91cf\u673a \u5b9e\u8df5 \u6587\u672c\u5206\u7c7b"},"content":{"rendered":"<h2>\u4e00\u3001HanLP \u57fa\u4e8eSVM\u652f\u6301\u5411\u91cf\u673a\u5206\u7c7b\u5668<\/h2>\n<p>\u4e0a\u7bc7\u6587\u7ae0\u901a\u8fc7\u6734\u7d20\u8d1d\u53f6\u65af\u6587\u672c\u5206\u7c7b\u5668\uff0c\u8bad\u7ec3\u6d4b\u8bd5\u4e86 <code>\u641c\u72d7\u6587\u672c\u5206\u7c7b\u8bed\u6599\u5e93\u8ff7\u4f60\u7248<\/code> \uff0c\u672c\u7bc7\u7ee7\u7eed\u6d4b\u8bd5SVM\u652f\u6301\u5411\u91cf\u673a\u5206\u7c7b\u5668\u3002<\/p>\n<p>\u7531\u4e8e<code>HanLP<\/code> \u5b98\u65b9\u7ed9\u51fa\u7684 <code>SVM<\/code> \u5206\u7c7b\u5668\u4f9d\u8d56\u4e86\u7b2c\u4e09\u65b9\u5e93\uff0c\u6ca1\u6709\u96c6\u6210\u5728\u4e3b\u9879\u76ee\u4e2d\uff0c\u9700\u8981\u62c9\u53d6 <code>text-classification-svm<\/code> \u9879\u76ee\uff1a<\/p>\n<blockquote>\n<p>git clone <a href=\"https:\/\/github.com\/hankcs\/text-classification-svm.git\">https:\/\/github.com\/hankcs\/text-classification-svm.git<\/a><\/p>\n<\/blockquote>\n<p>\u62c9\u53d6\u9879\u76ee\u540e\uff0c\u53d1\u73b0 pom \u4f9d\u8d56\u7684 <code>hanlp<\/code> \u7248\u672c\u662f <code>portable-1.5.2<\/code>\uff0c\u524d\u9762\u6211\u4eec\u7528\u7684\u90fd\u662f<code>portable-1.8.3<\/code>\uff0c\u8fd9\u91cc\u5c31\u4fdd\u6301\u4fdd\u672c\u4e00\u81f4\uff0c\u5c06 pom \u4e2d <code>hannlp<\/code> \u7684\u7248\u672c\u5347\u7ea7\u4e3a <code>portable-1.8.3<\/code>\uff1a<\/p>\n<p><img src=\"https:\/\/lemon-guess.oss-cn-hangzhou.aliyuncs.com\/img\/640-20251107140037921.png\" alt=\"\u56fe\u7247\" \/><\/p>\n<p>\u5347\u7ea7\u540e\u4f1a\u6709\u4e24\u70b9\u4e0d\u517c\u5bb9\uff0c\u9700\u8981\u4fee\u6539 <code>LinearSVMClassifier<\/code> \u7c7b\uff1a<\/p>\n<p><img src=\"https:\/\/lemon-guess.oss-cn-hangzhou.aliyuncs.com\/img\/640-20251107140037972.png\" alt=\"\u56fe\u7247\" \/><\/p>\n<p>\u7136\u540e\u5c06\u8be5\u9879\u76ee\u6253\u5305\u5230\u672c\u5730 <code>Maven<\/code> \u4e2d\u65b9\u4fbf\u5176\u4ed6\u9879\u76ee\u4f7f\u7528\uff1a<\/p>\n<pre><code>mvn clean install -DskipTests<\/code><\/pre>\n<p>\u4e0b\u9762\u5728\u81ea\u5df1\u7684 <code>Maven<\/code> \u9879\u76ee\u4e2d\u6dfb\u52a0\u4f9d\u8d56\uff1a<\/p>\n<pre><code>&lt;!-- svm \u5206\u7c7b\u5668\u6269\u5c55\u5305--&gt;\n&lt;dependency&gt;\n   &lt;groupId&gt;com.hankcs.nlp&lt;\/groupId&gt;\n   &lt;artifactId&gt;text-classification&lt;\/artifactId&gt;\n   &lt;version&gt;1.0.1&lt;\/version&gt;\n&lt;\/dependency&gt;\n&lt;!-- svm \u4f9d\u8d56 liblinear--&gt;\n&lt;dependency&gt;\n   &lt;groupId&gt;de.bwaldvogel&lt;\/groupId&gt;\n   &lt;artifactId&gt;liblinear&lt;\/artifactId&gt;\n   &lt;version&gt;1.95&lt;\/version&gt;\n&lt;\/dependency&gt;<\/code><\/pre>\n<p>\u6709\u5173\u4e8e HanLP \u73af\u5883\u7684\u642d\u5efa\uff0c\u53ef\u4ee5\u53c2\u8003\u4e0b\u9762\u8fd9\u7bc7\u6587\u7ae0\uff1a<\/p>\n<blockquote>\n<p><a href=\"https:\/\/xiaobichao.blog.csdn.net\/article\/details\/128271909\">https:\/\/xiaobichao.blog.csdn.net\/article\/details\/128271909<\/a><\/p>\n<\/blockquote>\n<p><code>SVM<\/code> \u6587\u672c\u5206\u7c7b\u5668\uff0c\u5229\u7528<code>liblinear<\/code>\u5b9e\u73b0<code>HanLP<\/code>\u4e2d\u7684\u6587\u672c\u5206\u7c7b\u63a5\u53e3\uff0c\u56e0\u6b64\u5bf9\u9884\u6599\u5e93\u7684\u8981\u6c42\u8fd8\u662f \u5c06\u6570\u636e\u96c6\u6839\u636e\u5206\u7c7b\u653e\u5230\u4e0d\u540c\u7684\u76ee\u5f55\u4e2d\uff1a<\/p>\n<p><img src=\"https:\/\/lemon-guess.oss-cn-hangzhou.aliyuncs.com\/img\/640-20251107140038017.png\" alt=\"\u56fe\u7247\" \/><\/p>\n<p>\u4e0b\u9762\u8fd8\u662f\u4f7f\u7528 <code>\u641c\u72d7\u6587\u672c\u5206\u7c7b\u8bed\u6599\u5e93\u8ff7\u4f60\u7248<\/code> \u8fdb\u884c\u6d4b\u8bd5\u3002<\/p>\n<h3>\u51c6\u5907\u8bed\u6599\u5e93<\/h3>\n<p>\u4e0b\u8f7d\u6570\u636e\u96c6\uff1a<\/p>\n<blockquote>\n<p><a href=\"http:\/\/file.hankcs.com\/corpus\/sogou-text-classification-corpus-mini.zip\">http:\/\/file.hankcs.com\/corpus\/sogou-text-classification-corpus-mini.zip<\/a><\/p>\n<\/blockquote>\n<p>\u4e0b\u8f7d\u52a0\u538b\u540e\u6570\u636e\u683c\u5f0f\uff1a<\/p>\n<p><img src=\"https:\/\/lemon-guess.oss-cn-hangzhou.aliyuncs.com\/img\/640-20251107140038065.png\" alt=\"\u56fe\u7247\" \/><\/p>\n<p>\u6587\u672c\u6837\u4f8b\uff1a<\/p>\n<p><img src=\"https:\/\/lemon-guess.oss-cn-hangzhou.aliyuncs.com\/img\/640-20251107140038379.png\" alt=\"\u56fe\u7247\" \/><\/p>\n<h3>\u8bad\u7ec3\u6570\u636e<\/h3>\n<pre><code>public class ClassifyTrain {\n\n    public static void main(String[] args) throws IOException {\n        \/\/\u8bed\u6599\u5e93\u7684\u5730\u5740\n        String dataPath = &quot;F:\/bigdata\/hanlp\/sogou-text-classification-corpus-mini\/\u641c\u72d7\u6587\u672c\u5206\u7c7b\u8bed\u6599\u5e93\u8ff7\u4f60\u7248&quot;;\n        \/\/\u6a21\u578b\u4fdd\u5b58\u8def\u5f84\n        String modelPath = &quot;F:\/bigdata\/hanlp\/sogou-text-classification-corpus-mini\/svm-classification-model.ser&quot;;\n        \/\/\u8bad\u7ec3\u6570\u636e\n        trainData(dataPath, modelPath);\n    }\n\n    private static void trainData(String dataPath, String modelPath) throws IOException {\n        File corpusFolder = new File(dataPath);\n        if (!corpusFolder.exists() || !corpusFolder.isDirectory())  {\n            System.err.println(&quot;\u6ca1\u6709\u6587\u672c\u5206\u7c7b\u8bed\u6599&quot;);\n            return;\n        }\n        \/\/ FileDataSet\u7701\u5185\u5b58\uff0c\u53ef\u52a0\u8f7d\u5927\u89c4\u6a21\u6570\u636e\u96c6\uff0c\u652f\u6301\u4e0d\u540c\u7684ITokenizer\uff0c\u8be6\u89c1\u6e90\u7801\u4e2d\u7684\u6587\u6863\n        \/\/ \u4f7f\u7528\u524d90% \u7684\u6570\u636e\u4f5c\u4e3a\u8bad\u7ec3\u96c6\n        IDataSet trainingCorpus = new FileDataSet()\n                .setTokenizer(new HanLPTokenizer())\n                .load(dataPath, &quot;UTF-8&quot;, 0.9);\n        \/\/ \u521b\u5efaSVM\u5206\u7c7b\u5668\n        IClassifier classifier = new LinearSVMClassifier();\n        \/\/ \u8bad\u7ec3\u6570\u636e\n        classifier.train(trainingCorpus);\n        \/\/ \u83b7\u53d6\u8bad\u7ec3\u6a21\u578b\n        AbstractModel model = classifier.getModel();\n        \/\/ \u4f7f\u7528\u540e10% \u7684\u6570\u636e\u4f5c\u4e3a\u6d4b\u8bd5\u96c6\n        IDataSet testingCorpus = new MemoryDataSet(model)\n                .load(dataPath, &quot;UTF-8&quot;, -0.1);\n        \/\/ \u8ba1\u7b97\u51c6\u786e\u7387\n        FMeasure result = Evaluator.evaluate(classifier, testingCorpus);\n        System.out.println(&quot;\u6d4b\u8bd5\u96c6\u51c6\u786e\u5ea6\uff1a&quot;);\n        System.out.println(result);\n        \/\/ \u4fdd\u5b58\u6a21\u578b\n        IOUtil.saveObjectTo(model, modelPath);\n    }\n\n}<\/code><\/pre>\n<p>\u67e5\u770b\u8bad\u7ec3\u65e5\u5fd7\uff1a<\/p>\n<p><img src=\"https:\/\/lemon-guess.oss-cn-hangzhou.aliyuncs.com\/img\/640-20251107140038454.png\" alt=\"\u56fe\u7247\" \/><\/p>\n<p>\u67e5\u770b\u8bad\u7ec3\u6a21\u578b\uff1a<\/p>\n<p><img src=\"https:\/\/lemon-guess.oss-cn-hangzhou.aliyuncs.com\/img\/640-20251107140038505.png\" alt=\"\u56fe\u7247\" \/><\/p>\n<h3>\u6d4b\u8bd5\u6a21\u578b<\/h3>\n<pre><code>public class TestClassify {\n\n    public static void main(String[] args) {\n        String modelPath = &quot;F:\/bigdata\/hanlp\/sogou-text-classification-corpus-mini\/svm-classification-model.ser&quot;;\n        testModel(modelPath);\n    }\n\n    private static void testModel(String modelPath){\n        LinearSVMModel model = (LinearSVMModel) IOUtil.readObjectFrom(modelPath);\n        IClassifier classifier = new LinearSVMClassifier(model);\n        \/\/ \u6d4b\u8bd5\u5206\u7c7b\n        String text1 = &quot;\u7814\u7a76\u751f\u8003\u5f55\u6a21\u5f0f\u4e9f\u5f85\u8fdb\u4e00\u6b65\u4e13\u4e1a\u5316&quot;;\n        System.out.printf(&quot;\u300a%s\u300b \u5c5e\u4e8e\u5206\u7c7b \u3010%s\u3011\\n&quot;, text1, classifier.classify(text1));\n\n        String text2 = &quot;C\u7f57\u83b72018\u73af\u7403\u8db3\u7403\u5956\u6700\u4f73\u7403\u5458 \u5fb7\u5c1a\u8363\u81ba\u6700\u4f73\u6559\u7ec3&quot;;\n        System.out.printf(&quot;\u300a%s\u300b \u5c5e\u4e8e\u5206\u7c7b \u3010%s\u3011\\n&quot;, text2, classifier.classify(text2));\n\n        String text3 = &quot;\u82f1\u56fd\u9020\u822a\u6bcd\u8017\u65f68\u5e74\u4ecd\u672a\u670d\u5f79 \u88ab\u4e2d\u56fd\u901f\u5ea6\u8fdc\u8fdc\u7529\u5728\u8eab\u540e&quot;;\n        System.out.printf(&quot;\u300a%s\u300b \u5c5e\u4e8e\u5206\u7c7b \u3010%s\u3011\\n&quot;, text3, classifier.classify(text3));\n\n        String text4 = &quot;\u5982\u679c\u771f\u60f3\u7528\u98df\u7269\u89e3\u538b,\u5efa\u8bae\u53ef\u4ee5\u98df\u7528\u71d5\u9ea6&quot;;\n        System.out.printf(&quot;\u300a%s\u300b \u5c5e\u4e8e\u5206\u7c7b \u3010%s\u3011\\n&quot;, text4, classifier.classify(text4));\n\n    }\n\n}<\/code><\/pre>\n<p>\u6d4b\u8bd5\u7ed3\u679c\uff1a<\/p>\n<p><img src=\"https:\/\/lemon-guess.oss-cn-hangzhou.aliyuncs.com\/img\/640-20251107140038550.png\" alt=\"\u56fe\u7247\" \/><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u4e00\u3001HanLP \u57fa\u4e8eSVM\u652f\u6301\u5411\u91cf\u673a\u5206\u7c7b\u5668 \u4e0a\u7bc7\u6587\u7ae0\u901a\u8fc7\u6734\u7d20\u8d1d\u53f6\u65af\u6587\u672c\u5206\u7c7b\u5668\uff0c\u8bad\u7ec3\u6d4b\u8bd5\u4e86 \u641c\u72d7\u6587\u672c\u5206\u7c7b\u8bed\u6599\u5e93\u8ff7\u4f60\u7248 \uff0c\u672c\u7bc7\u7ee7\u7eed\u6d4b\u8bd5SVM   \u2026 &#8230;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ngg_post_thumbnail":0},"categories":[225],"tags":[],"_links":{"self":[{"href":"\/index.php?rest_route=\/wp\/v2\/posts\/9470"}],"collection":[{"href":"\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=9470"}],"version-history":[{"count":1,"href":"\/index.php?rest_route=\/wp\/v2\/posts\/9470\/revisions"}],"predecessor-version":[{"id":9471,"href":"\/index.php?rest_route=\/wp\/v2\/posts\/9470\/revisions\/9471"}],"wp:attachment":[{"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=9470"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=9470"},{"taxonomy":"post_tag","embeddable":true,"href":"\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=9470"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}