Rainfall Forecasting for Raising the Yield Production Using Machine Learning Algorithms

摘要

In India, agriculture plays a major role in majority of lives. But the farmers are facing plenty of problems like rainfall shortage, animals mischievous, soil moisturizer, that causes poor production of yield. More than 65% of the crops yields in Tamil Nadu mainly depend upon seasonal rainfall. A modern innovation in Information and Communication Technology for agricultural sector is a fascinating research domain to predict/forecast the productivity of cultivation with respect to rainfall. The crisis of productivity forecasting is a key research area and can be done based on the historical dataset. Machine learning algorithms like random forest, support vector can be used for the prediction of rainfall or cultivation yield. The objective of this paper is to use different machine learning algorithms to forecast rainfall for raising the yield production in south Tamil Nadu particularly in Madurai. The experiment is conducted with monsoon dataset 4116 rows and 19 attributes and result is evaluated by using several metrics-Root mean square error (RMSE), mean square error (MSE) and R2 score. This research work also concluded Recurrent Neural Network works better than other types of machine learning algorithm namely multiple linear regression, Support vector regression, and Radom forest for forecasting rainfall to raise yield production.