Year: 2023
Type: article
Abstract: In the course of this research, we conducted an examination that specifically aimed to predict hourly rainfall. The approach we adopted for this analysis treated the problem as a binary classification task, where we categorized rainfall events into two groups: "rainy" (considered the positive class) and "non-rainy" (designated as the negative class). To forecast rainfall conditions for the upcoming hour, we utilized various independent climatic variables from the current hour. The data source we utilized was the CST weather station, which provided us with records of eight hourly weather parameters. To make our predictions, we employed a widely used machine learning model known as the Support Vector Machine(SVM), and this yielded an accuracy rate of 78%. Key Words: rainfall, SVM, prediction
DOI: 10.55041/ijsrem26209
Author: Manoj Chhetri
Institutions: College of Science and Technology, Royal University of Bhutan
Language: English
Subfield: Environmental Engineering
Field: Environmental Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4387784741