/clustering

Comparison among K-Means, DBSCAN and Density Peak

Primary LanguageMATLAB

Clustering

三种聚类算法的比较,在不同数据分布的数据集上测试这三个算法。

Dataset Download

Click this to get the dataset.

Description

该项目包含3个聚类算法,分别是K-Means, DBSCAN, DPeak算法,其中datasets文件包含了6个数据集,它们的散点图对应如下:

Results

>>kmeans('datasets/1/', 7, 0.01)
>>dbscan('datasets/1/', 4, 6)
>>dpeak_auto('datasets/1/', 4, 'gaussian',7)

1

>>kmeans('datasets/2/', 5, 0.01)
>>dbscan('datasets/2/', 10, 3)
>>dpeak_auto('datasets/2/', 3, 'gaussian',5)

2

>>kmeans('datasets/3/', 5, 0.0275)
>>dbscan('datasets/3/', 8, 0.023)
>>dpeak_auto('datasets/3/', 4, 'gaussian',5)

3

>>kmeans('datasets/iris/', 3, 0.01)
>>dbscan('datasets/iris/', 1, 0.078)
>>dpeak_auto('datasets/iris/', 4, 'gaussian',3)

iris

>>kmeans('datasets/wine/', 3, 0.01)
>>dbscan('datasets/wine/', 1, 0.13)
>>dpeak_auto('datasets/wine/', 5.64, 'gaussian',3)

wine

>>kmeans('datasets/spiral/', 3, 0.05)
>>dbscan('datasets/spiral/', 15, 3.5689)
>>dpeak_manual('datasets/spiral/', 5.64, 'gaussian')

spiral