Prediction of rainfall using data mining techniques

This thesis report is submitted in partial fulfilment of the requirements for the degree of Bachelor of Science in Computer Science and Engineering, 2017.Cataloged from PDF version of thesis report.Includes bibliographical references (pages 36-37).The research for this paper concentrates on finding inter-relations between various climatic indices and predict precipitation consequently. And since rainfall is the prominent reason behind flood, our study can aid immensely in predicting flood and designing a proper risk management system. Flood has been a major hindrance in the path of development for Bangladesh. Being a riverine country, flood occurs in Bangladesh almost every other year. Predicting flood accurately can help us in developing our economy. Our study shows how the climatic parameters (SOI,El Nino) are responsible for major rainfall in Bangladesh. Though many other researches on predicting rainfall have been conducted using other climatic factors, the southern oscillation index and the El nino 3.4 show stronger correlation with rainfall in our country than the others. For establishing a relationship among rainfall ,SOI and El Nino , we have applied Data Mining technique. The specific data mining algorithms that we have implemented in our paper are K-clustering, Decision tree and Regression model. The outputs of these algorithms give us a straightforward relationship between rainfall and the input parameters. Implementing our method on the dataset of rainfall for the past couple of years, our estimated rainfall is almost the same as the actual ones of those years. So in designing a feasible rainfall prediction model for Bangladesh, our work can play a significant role due to its high efficiency.Fahmida Tasnim PremaSharmin AhmedMd. Rifat IslamB. Computer Science and Engineerin