Support vector clustering of time series data with alignment kernels

• We cluster time series with a support vector clustering algorithm in a novel approach by using a Triangular Alignment Kernel.
• We thus cluster time series without determining the number of clusters or their shape in advance.
• We show that the quality of our results is competitive with other clustering approaches using the same dataset.

Time series clustering is an important data mining topic and a challenging task due to the sequences’ potentially very complex structures. In the present study we experimentally investigate the combination of support vector clustering with a triangular alignment kernel by evaluating it on an artificial time series benchmark dataset. The experiments lead to meaningful segmentations of the data, thereby providing an example that clustering time series with specific kernels is possible without pre-processing of the data. We compare our approach and the results and learn that the clustering quality is competitive when compared to other approaches.