/MPM_EDA

Java实现的一种层次聚类算法实验代码, 包含K-means实现作为benchmark

Primary LanguageJava

MPM_EDA_2013_written

Java实现的一种层次聚类算法，源码包中同时包含经典的K-Means算法作为对照组。 在iris和yeast等公开数据集上，该算法在参数调教合适时性能能够超越K-Means和另一种层次聚类算法——BIRTH。

程序发布版(jar)运行方法如下所示，具体详情请阅读源码~

Manual of the MPM-EDA.jar

The MPM-EDA.jar provides an implement of the MPM-EDA agorithm, and an implement of standrad K-meams as benchmark. Before using this jar package, please make sure the JDK 7 has been correctly installed in your computer.

A simple demo is shown in the Fig. 1. This demo program will employ MPM-EDA algorithm on the yeast data set, and detailed results of the clustering will be printed.

Alt text

The format of your input data should be accordance with the data files we provide (iris, seeds and yeast. You can find the in the .rar file). Notice that if you run the algorithm on seeds data set, please remove the sentence of “System.setProperty("java.util.Arrays.useLegacyMergeSort", "true");”

To run the K-means algorithm, see the program shown in Fig. 2. Alt text