A Comparative Analysis of the Machine Learning Model for Rainfall Prediction in Cavite Province, Philippines
Pitz Gerald G. Lagrazon
a
,
Jennifer Edytha E. Japor
b
,
Renato R. Maaliw
a
,
Julie Ann B. Susa
a
,
Maria Rossana D. De Veluz
a
,
Ace C. Lagman
c
,
Manuel B. Garcia
d
,
Arnold B. Platon
e
a College of Engineering, Southern Luzon State University, Quezon, Philippines
b Southern Luzon State University, Quezon, Philippines
c Information Technology Dept., FEU Institute of Technology, Manila, Philippines
d Educational Innovation and Technology Hub, FEU Institute of Technology, Manila, Philippines
e Computer Studies Department, Bicol University Polangui, Albay, Philippines
Abstract: Rainfall is crucial for flood prevention and comprehending the correlation between rainfall and flooding. Cavite province in the Philippines is vulnerable to flooding caused by heavy rainfall and climate change impacts. Early detection of flooding through early warning systems can prevent excessive damage loss and potentially save lives. It can also provide major savings in terms of monetary benefit and increased interagency coordination for rapid decision-making. Machine learning is an important tool for predicting rainfall which can be used to predict rainfall in the province. The objective of this study is to conduct a comparative analysis of various models for predicting daily rainfall, using relevant atmospheric features such as maximum, minimum, and mean temperature, relative humidity, wind speed, wind direction, cloud cover, pressure, and evaporation. The study seeks to identify the most effective model for accurately predicting rainfall in the Cavite Province to benefit the local community. Among the five machine learning models evaluated, the Gaussian Process Regression model demonstrated the highest accuracy in predicting daily rainfall. The findings of this study can be leveraged to mitigate the damage caused by flooding in the Cavite Province and serve as a useful reference for similar studies in other regions prone to flooding.