Determining relative humidity in Surakarta city using semiparametric regression based on Fourier series penalized least square

Abstract

This study discusses the use of a semiparametric regression model approach based on Fourier series penalized least square estimator to determine the relationship between relative humidity and dew point in Surakarta city of Indonesia. The proposed method can address complex climate data patterns. A dataset of 100 observations was analyzed under three training data scenarios, for sample sizes of 𝑛 = 70, 𝑛 = 80, and 𝑛 = 90. It yields the optimal Fourier coefficients of 2, 2, and 2, and smoothing parameter values of 0.018, 0.0124, and 0.039, with minimum generalized cross validation values of 4.410572, 4.191036, and 5.989094. The results of this study show that the proposed method provided good performance for prediction purpose with Mean Absolute Percentage Error (MAPE) values of 2.471245, 2.270436, and 2.93358. This means that category of the proposed method is highly accurate for prediction purpose. These results underline the ability of the proposed method to capture the inverse relationship between humidity and dew point. In other words, based on these results, this study highlights the effectiveness of the semiparametric regression model approach based on Fourier series penalized least square estimator in dynamic scenarios and recommends its application to other climate variables or regions to further evaluate its adaptability and resilience.