共 82 篇文章 |
|
在eclipse下使用java调用weka.->浏览weka所在目录,将weka.jar添加进来,然后单击ok。import java.io.import weka.classifiers.import weka.core.File inputFile = new File("D://Program Files//Weka-3-6//data//cpu.with.vendor.arff");if(m_classifier.classifyInstance(instancesTest.instance(i))==instancesTest.instance(i... 阅272 转11 评0 公众公开 14-07-07 19:54 |
Weka中实现了weka.core.OptionHandler接口,这个接口为比如classifiers,clusterers,filers等提供了设置,获取参数的功能,函数如下:一个clusterer建立与建立一个分类器的方式相似,只是不是使用buildClassifier(Instances)方法,它使用buildClusterer(Instances),下面的代码段展示了如何用EM clusterer使用最多100次迭代的方法。l 顺序调... 阅300 转7 评0 公众公开 14-05-31 14:46 |
Weka开发[48]——用Weka文本分类。java weka.core.converters.默认情况下,非数值型属性会以NOMINAL(离散型)属性方式导入,这对于文本型数据是不太合适的,特别是有人想在后面使用StringToWordVector过滤器时,要将属性指定为STRING类型,可以对数据运行NominalToString过滤器(在weka.filters.unsupervised.attribute包中),并指定属性下标或... 阅3091 转8 评0 公众公开 13-07-22 21:22 |
protected Instances metaFormat(Instances instances) throws Exception {Instances metaFormat;Instances是level 0的训练样本集,现在要加入一部分属性用以保存后来的分类结果,如果m_BaseFormat类别属性是连续值,那么就加入m_Classifiers个属性,如果是离散值,每次要加入level 0类别属性取值个数个属性,最后加入metaFormat的类别属性。pr... 阅463 转4 评0 公众公开 13-07-22 21:22 |
结果的意思大致如下Total population就是有多少个样本,这里是150个,Target attribute是指感兴趣的属性是哪个,这里是class属性,Target value是感兴趣的属性中感兴趣的值,这里是Iris-setosa,一共有50个样本属于这个类别值,接下来是支持次数是50,支持度是33%。double[] probs = new double[inst.attributeStats(m_target)double[] subsetM... 阅131 转2 评0 公众公开 13-07-22 21:21 |
public void buildStructure(BayesNet bayesNet, Instances instances) throws Exception {public void addParent(int nParent, Instances _Instances) {int nOrder[] = new int[instances.numAttributes()];nCounts[numValues * ((int) iCPT) (int) instance.value(nNode)] ;public void updateClassifier(BayesNet bayesNet, Instance instan... 阅358 转5 评0 公众公开 13-07-22 21:21 |
m_modelNormal = new double[m_wrappedClusterer.numberOfClusters()][data.double[][] weights = new double[m_wrappedClusterer.public void addValue(double data, double weight) {double diff = m_modelNormal[clusterIndex[i]][j][0]double[] wghts = new double[m_wrappedClusterer.numberOfClusters()];public double getProbability(d... 阅607 转2 评0 公众公开 13-07-22 21:21 |
DataObject dataObject = dataObjectForName(DataObject dataObject = (DataObject) iterator.next();public List coreDistance(int minPoints, double epsilon, DataObject dataObject) {List list = k_nextNeighbourQuery(minPoints, epsilon, dataObject);dataObject.setReachabilityDistance(DataObject.update(seeds, epsilonRange_List, ... 阅402 转3 评0 公众公开 13-07-22 21:21 |
database.insert(dataObject);public void insert(DataObject dataObject) {DataObject dataObject = (DataObject) iterator.next();if (dataObject.getClusterLabel() == DataObject.dataObject.setClusterLabel(DataObject.DataObject seedListDataObject = (DataObject) seedList.get(i);DataObject seedListDataObject = (DataObject) seed... 阅938 转3 评0 公众公开 13-07-22 21:21 |
m_ClusterCentroids.add(m_instances.instance(firstI));protected void updateMinDistance(double[] minDistance, boolean[] selected, Instances data, Instance center) {updateMinDistance(minDistance, selected, m_instances, m_instances.m_instances = new Instances(m_instances, 0);protected double distance(Instance first, Insta... 阅525 转3 评0 公众公开 13-07-22 21:21 |