JAVA 使用weka进行KNN预测

偏执的太偏执、 2022-01-21 23:07 759阅读 0赞

## maven ##

<dependency>
          <groupId>nz.ac.waikato.cms.weka</groupId>
          <artifactId>weka-stable</artifactId>
          <version>3.8.3</version>
        </dependency>

以线性查询为例，别的实现方法替换对应的实现类即可

/**
         *
         * @param sourceIns 表示你在哪个数据集中找邻居
         * @param target 找邻居的样本
         * @param kNN 多少邻居
         * @return k个邻居
         */
        public static Instances getKNeighbour(Instances sourceIns, Instance target, int kNN)
        {
            Instances neighbours = null;
            try{
                EuclideanDistance dfunc = new EuclideanDistance();
                LinearNNSearch lnn = new LinearNNSearch();
                lnn.setDistanceFunction(dfunc);
                lnn.setInstances( sourceIns );
                lnn.addInstanceInfo(target);
                neighbours = lnn.kNearestNeighbours(target, kNN);
            }catch(Exception e){
                System.out.println("Util.buildKdTree(LinearNNSearch m_NNSearch,Instance mean, int m_kNN) is wrong!");
            }
            return neighbours;
        }

把arff文件读取成Instances

private static Instances getFileInstances(String fileName) throws Exception {
            ConverterUtils.DataSource frData = new ConverterUtils.DataSource( fileName );
            return frData.getDataSet();
        }

跑一个例子：

public static void main(String[] args) throws Exception {
        String path = "D:\\poi_test.arff";
        Instances instances = getFileInstances(path);
        System.out.println("instances:" + instances);
        ArrayList<Attribute> atts = new ArrayList<>();
        atts.add(new Attribute("x", Attribute.NUMERIC));
        atts.add(new Attribute("y", Attribute.NUMERIC));
        atts.add(new Attribute("class", Attribute.NUMERIC));
        Instances df = new Instances("predictData", atts, 0);
        df.setClassIndex(df.numAttributes() - 1);
        Instance target = new DenseInstance(2);
        target.setValue(0, 116.41);
        target.setValue(1, 39.95);
        System.out.println("target:" + target);
        df.add(target);
        System.out.println("=============knn=================:");
        Instances result = getKNeighbour(instances, target, 3);
        System.out.println("result:" + result);
    }

数据：

@relation distance 
    
    @attribute x NUMERIC
    @attribute y NUMERIC
    @attribute class NUMERIC
    
    @data
    116.415668,39.953726,0
    116.438614,39.930745,0
    116.426367,39.928047,0