Time series is an important class of temporal data objects and it can be easily obtained from scientific and financial applications. The nature time series data includes: large in data size, high dimensionality and necessary to update continuously. The increasing use of time series data has initiated a great deal of research and developing attempts in the field of data mining. Classification of time series data has a wide range of applications and has attracted researches from a wide range of discipline. In this paper the classical discriminant analysis is modified using the principal component analysis (PCA) to overcome the large dimensionality. The PCA modification can reduce the size of the data and improve the efficiency and accuracy. The new method is investigated using a simulation study to classify the linear AR(2) model and the bilinear BL(1,0,1,1) model. The results of our investigation show that the designed algorithm has a significant rate of correct classification especially if it is compared with the other methods. The PCA modification method is also applied to a real set of time series data and gave a superior rate of correct classification.
Digital Object Identifier (DOI)
M. Gabr, M. and M. Fatehy, L.
"Time Series Classification,"
Journal of Statistics Applications & Probability: Vol. 2
, Article 5.
Available at: https://dc.naturalspublishing.com/jsap/vol2/iss2/5