java中libsvm用于分类判断

最新推荐文章于 2021-02-27 04:56:24 发布

小_秋

最新推荐文章于 2021-02-27 04:56:24 发布

阅读量192

点赞数

分类专栏： svm 文章标签： libSVM java

本文链接：https://blog.csdn.net/u011878544/article/details/49174463

版权

svm 专栏收录该内容

1 篇文章 0 订阅

订阅专栏

1. 批量分类

String[] trainArg = { "trainingSet.txt", "train_model" };
svm_train.main(trainArg);
String[] testArg = { "trainingSet.txt", "train_model","predictLabel.txt" };
svm_predict.main(testArg);

2.单个样本实时分类

假设model已经按照1中的方式训练好了。

public int getModelKind(double []feature) {
	int modelKind = 0;
	svm_node[] arrayOfsvm_node = new svm_node[feature.length];
	for (int i = 0; i < arrayOfsvm_node.length; i++) {
	<span style="white-space:pre">	</span>arrayOfsvm_node[i] = new svm_node();
		arrayOfsvm_node[i].index = i;
		arrayOfsvm_node[i].value = feature[i];
	}
	try {
	<span style="white-space:pre">	</span>svm_model localsvm_model = svm.svm_load_model("train_model");
		double kind = svm.svm_predict(localsvm_model, arrayOfsvm_node);
		System.out.println("kind1: " + kind);
		modelKind = (int) kind;
	} catch (IOException e) {
		// TODO Auto-generated catch block
		e.printStackTrace();
	}

	return modelKind;
}