<![CDATA[Corn is an important food crop commodity that plays a significant role in the agricultural sector and regional economy. Efforts to increase corn production in Gorontalo have become one of the programs initiated by the Ministry of Agriculture to support Indonesia's corn exports. However, corn productivity is influenced by various factors, one of which is temperature variation resulting from climate change. This study aims to predict weekly maximum temperatures as a basis for determining the optimal planting time for corn by comparing the performance of the \textit{Gaussian Process Regression} (GPR) method using four kernel functions (\textit{Periodic}, \textit{Matern}, \textit{Radial Basis Function}, and \textit{Rational Quadratic}) and the \textit{Support Vector Regression} (SVR) method optimized using \textit{Particle Swarm Optimization} (PSO). The data used in this study consist of weekly maximum temperature observations from 2023 to 2024 obtained from the Gorontalo Climatology Station. The results indicate that the GPR method achieved the best performance, yielding a \textit{Mean Absolute Percentage Error} (MAPE) of 2.17\%, while the PSO-optimized SVR method produced a MAPE of 2.30\%. Based on the forecasting results, the optimal corn planting period within the next 30 weeks is between the sixth and eighth weeks, as the critical growth phase of the crop is expected to occur under the most stable temperature conditions and within the optimal temperature range for corn growth. ]]>
Copyrights © 2026
