Linear Regression Analysis Using Log Transformation Model for Rainfall Data in Water Resources Management Krueng Pase, Aceh, Indonesia

Abstract

Climate changes are one crucial factor that influenced water availability at one location since they affected the environmental, social, and agricultural systems. The study observed the agent factors that influenced the rainfall changes at Krueng Pasee Aceh watershed, Indonesia. The method used in this research is a linear regression with a log transformation approach on predictor variables. The data used in this study consisted of rainfall, a total of rainy days, temperature, humidity, duration of irradiation, and wind speed in the period ranging from 1992 to 2020. Results showed that the agent factors had not distributed normally. The regression model produced after log transformation had met the classical assumptions and can be used to predict the rainfall at R-quare 24.61% with an RMSE value of 57.676. From all factors studied, the wind speed should be excluded. Further study is recommended to use the nonlinear method to improve the model for rainfall prediction.