Forecasting of Resin Yield and Number of Blazes of Naturally Regenerated Chir Pine (Pinus roxburghii Sargent) in Himachal Pradesh by Using Single Exponential Smoothing Method

Abstract

The present investigation entitled “Forecasting of resin yield and number of blazes of naturally regenerated chir pine (Pinus roxburghii Sargent) in Himachal Pradesh by using single exponential smoothing method” was carried out during the year 2020-2022. The secondary information was collected regarding the resin yield and number of resin blazes w.e.f. 2005 to 2022 from Himachal Pradesh State Forest Department. Exponential smoothing is a particular moving average technique applied to time series data and to produce smoothed data to make forecast. In exponential smoothing, one or more smoothing parameters are to be determined explicitly and those choices determine the weights assigned to the observations. Forecasting with the help of various linear and non-linear models is on the assumption that the series is stationary. Often time series is found to be non-stationary which means they are integrated and can be made stationary by differencing the time series. To check the stationarity of the number of blazes and resin yield, Augmented Dickey Fuller test was used. The results indicated that data was not approaching stationarity even after taking third difference. So, the prerequisite condition of ARIMA model is that the data should be stationary and if the time series data does not contain trend and seasonal components, Single Exponential Smoothing model was used. The Single Exponential Smoothing model was found to be best fit for the prediction of number of blazes and resin yield as per the high value of , low value of MAPE and Normalized BIC.