Prediction of rain fall through the moving averages model Prakash K. Kiran, Reddy B. Ravindhar1, Abbaiah R.2 Dept of Statistics &Maths, Agricultural College, ANGRAU, Aswaraopet, A.P. 1Scientist, (Stat & Maths), RARS, ANGRAU, Tirupathi, A.P 2Professor, Dept of Statistics, S.V. University, Tirupathi, A.P Online published on 26 February, 2013. Abstract Earlier authors proposed (Kiran Prakash et al 2010) a preliminary statistical analysis to study the behavior of the rainfall in three Mandals namely (1) Sathupally, (2) Vemsoor and (3) Aswaraopet. Further, they proposed a Markov Chain model for the rainfall data by considering ‘0’ as no-rainy day and ‘1’ as a rainy day and obtained Steady state solutions (Kiran Prakash et al 2010), for the above three Mandals. In this paper a new approach for prediction of rainfall in these three Mandals is considered using Moving Average[M.A] Model. A time series (Xt, tî T) is called time series data and daily rainfall can be considered as a time series. For this data a Moving Average process of order ‘p’ can be fitted and predictions can be made based on the fitted model by putting p = 3. To fit the model, we have used method of ‘Least Squares’ (Gupta and Kapoor 2005) and obtained Mean Square Error (M.S.E) (Gupta and Kapoor 2005, Medhi 1984). These M.S.E. of different Mandals of M.A. (3) Process are compared and conclusions are drawn based on the results obtained. Top Keywords Moving Average, Least Square, Mean Square Error. Top |