An Improved Prediction Model for Observing Rainfall Pattern Using Novel Spatio-Temporal Evolution by Comparing Linear Regression Over Logistic Regression

Abstract

The aim is to classify spatio-temporal evolution logic to predict rainfall in advance for improving people's awareness rate. Materials and Methods: The Geographical Data Sharing of India dataset used for training and testing of the proposed prediction model, which consists of 1200 pattern visualization with 8 attributes. Classification is performed by invoking Logistic Regression with a sample size =10 and Linear Regression with a sample size =10 with a G power of 80% and threshold 0.05%, CI 95% mean and standard deviation. Results and Discussion: In this work, the comparison of linear and logistic regression has classified and predicted the values from the rainfall data to generate accuracy with linear regression has higher accuracy (87.32%) comparison with Logistic Regression accuracy (80.35%). Conclusion: Prediction in rainfall patterns, linear regression consisting of classification to rainfall, has significantly been used to generate better accuracy than the Logistic Regression.