Feature Selection Using Efficient Independent Component Analysis (EICA) for Weather Data Analysis

Abstract

Finding the most pertinent features in the study of weather data analysis is an essential first step in creating accurate and effective models. Principal component analysis (PCA) and recursive feature elimination (RFE), two common feature selection techniques, can be computationally expensive and may not be appropriate for huge datasets. In this article, we provide an effective feature selection technique for weather data analysis termed Efficient Independent Component Analysis (EICA). The most crucial independent components in the dataset are found using EICA, a modified variant of Independent Component Analysis (ICA), which employs an effective model. We apply EICA to a dataset of actual weather conditions and assess how it performs in comparison to other feature selection techniques. The findings demonstrate that EICA beats PCA and ICA in terms of both accuracy and efficiency.