抽象的な

Ddbscan: a density detection dbscan algorithm in e-commerce sites evaluation

Jianhua Jiang, Haiyan Bian, Yumian Yang

To solve the problem of the uneven density data of DBSCAN algorithm, this paper proposes a density detection DBSCAN algorithm, which is named as DDBSCAN. Firstly, the density detection functions are designed as the evaluation standard of data density; secondly, high-dimensional data are classified into several partitions based on different density values; thirdly, Eps and MinPts parameters are set up in these partitions automatically; finally, the DBSCAN algorithm is applied to each partition respectively. Experimental results show that the proposed DDBSCAN algorithm is superior to the original DBSCAN in uneven density data clustering perspective.