Asian Journal of Probability and Statistics

The main reason for this study is to know the performance of BFTSC (Break for Time Series Components) and GFTSC (Group for Time Series Components) in identification of time series components using volatile simulated and empirical data. BFTSC was created to capture the trend, seasonal, cyclical and irregular components and presented them in a time series plot. While GFTSC was designed to capture all the four time series components together with the equations that produces each components of time series. BFAST (Break for Additive, Seasonal and Trend) only identifies trend and seasonal components while considering all other left over components as random, identification of trend and seasonal components alone is not enough to have a clear image of all the time series components in a time series data. Performance through evaluation using low and high volatile simulated and empirical data was conducted to evaluate the performance of both techniques. For yearly sample size of 8, 16 and 24 years were for small medium and large sample size. For the monthly data, 48, 96 and 144 months were used as small, medium and large sample size. Each of the sample size was replicated 100 times each. Finally, GFTSC and BFTSC performance was very good for large sample size with linear trend for both monthly and yearly data (approximately 100%). While the performance drops with highly volatile data such as trend with curve trend line (such as quadratic and cubic). These findings indicate that BFTSC and GFTSC can provide a better alternative to manual technique and BFAST for data associated with linear trend, hence BFTSC and GFTSC are recommended for public.