<![CDATA[Indonesia predominantly features a tropical climate across its entirety. With this mostly tropical climate, the country encounters minimal shifts in temperature but exhibits a wide array of rainfall variations. Rainfall patterns in Indonesia showcase significant diversity. These variations in rainfall hold substantial importance in mitigating risks linked to heavy rainfall, such as floods and landslides. Moreover, besides its role in disaster preparedness, rainfall data also holds practical value in sectors such as agriculture, transportation, and industry. By incorporating data mining classification techniques, the process of predicting rainfall in Indonesia can be greatly enhanced. In this study, daily climate data from Indonesia is harnessed, and the chosen method for classification is the random forest algorithm. This selection stems from its capability to generate accurate and consistent classification models without necessitating intricate adjustments of parameters. Furthermore, the Naïve Bayes method is also integrated due to its straightforward implementation and its capacity for simple probability modeling, which can be effectively applied across diverse classification data. The outcomes of this investigation suggest that the random forest algorithm surpasses the Naïve Bayes algorithm in terms of performance and accuracy when classifying climate datasets unique to Indonesia. The random forest algorithm attains an accuracy rate of 86.55%, whereas the Naïve Bayes algorithm lags at an accuracy rate of 36.61%. It is anticipated that these research findings can serve as a point of reference for subsequent scholarly inquiries and contribute to the ongoing monitoring of daily rainfall in Indonesia, thereby aiding in the prevention of natural disasters.]]>
Copyrights © 2024
