Concise Prediction of Early Floods in Variable Weather Forecasts using K Nearest Neighbor

Aim: In light of unpredictable weather forecasts, the goal of this research is to develop and assess a sophisticated K nearest Neighbor (KNN) based systematic prediction system for early flood detection in comparison to the Logistic Regression (LR) approach. Materials and Methods: The dataset was obtained from kaggle, a popular platform for collaborative projects, datasets, and machine learning and data science competitions. To improve the accuracy of the current research, a Logistic Regression was used, with a sample mean size of 20 sets. We are comparing 40 sets in total with K-Nearest Neighbor. The current study’s average accuracy has been calculated using the ClinCalc software appliance. The statistical analysis was presented using the SPSS software, and the Python programming language was used for implementation. Result: The K-Nearest Neighbor and Logistic Regression both achieved 98% and 85% accuracy after this investigation was finished. Statistical significance was shown by the significance value of p=0.000 (p<0.05) after conducting an independent samples T-Test analysis. Conclusion: The research adds to the Machine learning algorithm emphasizing that the current study combines the Logistic Regression with the K-Nearest Neighbor algorithm. Following the completion of the research, this highlights the potential of the K nearest Neighbor as a useful tool in this sector.