On the Data-Mining Oriented Methods for Clustering Time Series
面向数据挖掘的时间序列聚类方法研究

According to the characteristics of data mining,some improvements are made to the neural network fuzzy clustering algorithm FSART to make it more efficient and can perform incremental clustering. For the need of the clustering analysis of time series,a new fuzzy membership expression that can describe the modality similarity of vectors is proposed. The new fuzzy membership expression and the improved FSART algorithm are combined to implement the clustering analysis of the nonstationary time series.