Analysis of Computational and Statistical model For Time Series Prediction

Abstract

            Water is considered to be one of the most important substances on earth. All living things-plants, animals and human beings require water for their survival. Water is useful for day-to-day activities as well as  for global requirements . Thus, preserving water for various purposes becomes vital. There are so many ways by which water is collected. One of them being surface water which means water collected on the surface of the earth that includes  rivers, lakes etc., and most surface water comes from rainfall. Proper use of rainfall will be of great help for meeting the requirements. For a metro city like Chennai, this rainfall is a gift from God as water from other resources is very less. This research article discusses about the prediction by one of the smoothing time series technique Moving Average, computational data mining technique K-Nearest Neighbour (KNN) and the hybrid models KNN-Moving Average. Error measure Mean Absolute Percentage Error (MAPE) is used to test the validity of the model. The study makes an analysis of these models.