Affiliation of Author(s):计算机科学与技术学院/人工智能学院/软件学院
Journal:INTELLIGENT DATA ENGINEERING AND AUTOMATED LEARNING - IDEAL 2017
Key Words:Uncertain time series DTW with limited width UK-Means clustering ARI
Abstract:The probability density function represents the uncertainty of time series at each time point. In this paper, based on probability density function, we adopt the ULDTW distance for uncertain time series and apply it to the traditional UK-Means clustering. Combining the property that ULDTW distance has a one-to-many correspondence between time points in the matching process, we propose a 1ToNCenter calculation method replacing the traditional mean cluster-center calculation method to improve the accuracy of clustering results. Experiments show that the Adjusted Rand Index (ARI) of UKMeansULDTW clustering results have an obviously higher accuracy than the existing UK-Means algorithms in the high dimensional uncertain time series cases.
ISSN No.:0302-9743
Translation or Not:no
Date of Publication:2017-01-01
Co-author:Zhu, Xiaoping,Tang, Qijie
Correspondence Author:mazongmin
Date of Publication:2017-01-01