Big Data and Predictive Analytics: Time Series Forecasting for Improved Decision Making in Business Applications

Abstract

Predictive analytics is a significant aspect of big data analysis where large data sets are analyzed in order to give future predictions. Reliability and accuracy of the information generated through the use of predictive analysis models are crucial, particularly in the business environment. Many business organizations have progressed to integrate predictive analytics to their systems in order to get a competitive advantage in the business arena. Forecasting is critical to the successful execution of strategic as well as operational functions of an organization which further emphasizes the significance of the precision level of the forecasted information. Time series modeling and forecasting is a widely-used technique in monitoring and analyzing industrial processes to generate forecasted business data. This paper critically evaluates and analyses the moving average, exponential smoothing and linear regression predictive analytics models, which are widely used in time series forecasting, determining their reliability and accuracy with the use of business data. The accuracy of the time series forecasting models was further established through forecast error measurement statistics where it was determined that the linear regression model is a better fit for business data. The secular trend and the comparison of the actual data set with the forecasted data sets also helped determine the extent of accuracy of the predictive models.