Analisis Statistik Data Curah Hujan Harian Pada Das Siak Berdasarkan Aic (Akaike Information Criterion)

Abstract

Analysis of hydrological data is needed to determine the rainfall characteristics that suitable to space and time. Analysis of hydrological data can interpret the possibility of an upcoming event based on the data from the past. This interpretation uses probability distribution. The probability distribution that commonly used is Gamma Distribution, Exponential Distribution, Normal Distribution, Normal Log Distribution, Gumbel Distribution, and Log Person III Distribution. The purpose of this study was to choose the probability distribution based on Akaike Information Criterion (AIC) and to determine its compatibility with the Chi Square test and the Smirnov Kolmogorov test. The maximum likelihood value of each data distribution is evaluated by AIC value. The distribution that has the smallest AIC values were re-tested with Chi Square test and Smirnov Kolmogorov test. The results showed that in the Siak River Basin, Normal Distribution has the smallest AIC value with AIC value -2.203 for Buatan Station, -2.486 for Kandis Station, -2.939 for Pekanbaru Station and -3.184 for Petapahan Baru Station. However, Normal Distribution can not be accepted as the best distribution according to Chi Square test and Smirnov Kolmogorov test.
Keywords: Akaike Information Criterion (AIC), Probability Distribution, Chi Square test, Smirnov Kolmogorov test