FORECASTING SOFTWARE RELIABILITY USING ARIMA WITH ENSEMBLE EMPIRICAL MODE DECOMPOSITION

Yukun Bao, Dongbo Yi, and Junhua Chen

Keywords

Softwarereliability forecasting, Ensemble Empirical Mode Decomposition, Autoregressive Integrated Moving Average, time series forecasting

Abstract

In this study, a hybrid forecasting approach combining Autoregressive Integrated Moving Average (ARIMA) and Ensemble Empirical Mode Decomposition (EEMD) is proposed for software-reliability forecasting. This hybrid approach first decomposes the softwarereliability data series using EEMD into several subseries including a small number of intrinsic-mode functions and a residual. Then ARIMA is used to model each of the derived subseries and to generate a forecast for each. Finally, an ensemble forecast is obtained by summing up the forecasts for each subseries. A real softwarereliability data set obtained from the literature is used to evaluate the performance of the proposed approach. The experimental results demonstrate that the approach described here outperforms the other methods reported in the literature using the same experimental settings. Hence, the proposed approach is a valid and promising alternative for software-reliability forecasting.

Important Links:



Go Back