Badanie jakości klasyfikacji szeregów czasowych
Validation of time series clustering

Summary/Abstract: The aim of this paper is to present a quality index classification using Renyi’s entropy against known quality indicators grouping of multidimensional time series. The starting point was the empirical data. The division into groups was made by using Ward's agglomeration algorithm, k-means method’s and spectral clustering. The results were veri-fied using the selected indices of clustering validation. The proposed index seems to be promising but it needs to be verified for various distributions of time series.

• Details
• Contents