Comparative Study on Influence of Moon's Phases in Rainfall Prediction

Rainfall prediction is a routine part of meteorological observations. An essential work of this department includes keeping daily record of rainfall characteristics, variations, intensity, etc. for a particular region. Although, in our general life, we may not pay much attention to our day-to-day weather conditions, for instance, sunrise, humidity, rainfall, air pressure and others, but in terms of climatology, such weather events or their fluctuation can leave a great impact on the habitat, if they remain consistent in long run. For this reason, advance methods are being implemented to develop accurate weather prediction tools. Also, several researches are done in the area of climatology to confirm a presumed connection between our Earth's weather and unconventional, new unusual phenomenon that is noticed within the climate influencing atmospheric range. Our research is planned to analyze such a phenomenon. In our study, we constructed a Machine Learning Based Rainfall prediction Model with Moon's phases included as a feature to observe the its importance level in rainfall prediction as well as compare its value with other influencing factors. We have chosen Machine Learning approach to achieve desired accuracy, speed and efficiency than any contemporary manual engineering processes done for data analysis. We have incorporated two predictive algorithms, namely, Logistic Regression and Random Forest to enhance our Model's predictive potentiality. Since, lowering of computational speed, making erroneous calculations, increasing system processing risk are some of the difficulties of Machine Learning implementation with large dataset, we have utilized Feature Selection Technique to overcome them.