Categorical Time-Series Data Classification base on Sequential Pattern

Abstract

Rich kinds of time-series data exist in wide application domains. These data, like microarry data set, are usually hard to handle with common statistical methods. Inherently, there exist interesting correlation between the time-series data itself and some associative class label. The motivation of our research is to explore the issue of data classification based on time-series data. Although a number of methods have been proposed for solving the classification problem based on the well-known learning models like decision tree or neural network, they may not perform well in mining datasets with time sequence property like time-series gene expression data. In this paper, we propose a new data mining method, namely Classify-By- Sequence (CBS), for classifying large time-series datasets. The CBS method mainly utilizes the concept of sequential pattern mining and probabilistic reasoning. We designed two policies namely CBS-Class and CBS-All for predicting the class of new data instances. Finally, we evaluate the performance of CBS in comparison with other methods through several experiments. The experiments show that CBS achieves better performance in both of accuracy and execution efficiency.