Abstract:
The most significant climatic element which
impacts on agriculture sector is rainfall and rainfall
prediction becomes an important issue in agriculture
country like Myanmar. Collecting, storing and
processing of huge amount of climatic data (Big
Data) require high-performance analytical systems
running on distributed environments for accurate
prediction of weather. Traditional standard data
analytics algorithms need to be adapted to take
advantage of cloud computing models which provide
scalability and flexibility. In this paper, Multiple
Linear Regression which is an empirical, statistical
and mathematically mature method in data analysis
is applied in Rainfall Prediction. To prove
conventional Multiple Linear Regression work
efficiently in distributed environments, we propose a
parallel processing of Regression Model called
MapReduce-based Multiple Linear Regression (MRMLR).
Weekly Rainfall Prediction with the proposed
regression model using large scale weather data will
base on the QR Decomposition and Ordinary Least
Squares method adapted to MapReduce Framework.
Correlation-based Filter Approach by using
Symmetrical Uncertainty (SU) will be applied in
selecting correlated and relevant features for
improving the proposed regression model’s
prediction accuracy.
