2024-03-29T07:40:21Z
https://meral.edu.mm/oai
oai:meral.edu.mm:recid/4779
2021-12-13T02:27:19Z
1582963302567:1597824273898
user-ucsy
Big Data Analytics for Rainfall Prediction using MapReduce-Based Regression Model
Khine, Kyi Lai Lai
Nyunt, Thi Thi Soe
The most significant climatic element whichimpacts on agriculture sector is rainfall and rainfallprediction becomes an important issue in agriculturecountry like Myanmar. Collecting, storing andprocessing of huge amount of climatic data (BigData) require high-performance analytical systemsrunning on distributed environments for accurateprediction of weather. Traditional standard dataanalytics algorithms need to be adapted to takeadvantage of cloud computing models which providescalability and flexibility. In this paper, MultipleLinear Regression which is an empirical, statisticaland mathematically mature method in data analysisis applied in Rainfall Prediction. To proveconventional Multiple Linear Regression workefficiently in distributed environments, we propose aparallel processing of Regression Model calledMapReduce-based Multiple Linear Regression (MRMLR).Weekly Rainfall Prediction with the proposedregression model using large scale weather data willbase on the QR Decomposition and Ordinary LeastSquares method adapted to MapReduce Framework.Correlation-based Filter Approach by usingSymmetrical Uncertainty (SU) will be applied inselecting correlated and relevant features forimproving the proposed regression model’sprediction accuracy.
2017-02-16
http://hdl.handle.net/20.500.12678/0000004779
https://meral.edu.mm/records/4779