Volume No. 12 Issue No.: 4 Page No.: 519-528 April-June 2018

 

A STUDY TO EXTRACT ASSOCIATION RULE BETWEEN RAINFALL AND CLIMATIC INDICES USING DATA MINING TECHNIQUE

 

Mirzanur Rahman* and Geetali Das

Department of Information Technology, Gauhati University, Guwahati, Assam (INDIA)

 

Received on : September 29, 2017

 

ABSTRACT

 

Data mining is a process to discover important informationís from large amounts of data sets. Association rule mining is one of the areas of data mining to find out the relation between different sets of objects. The problem of deriving associations from data has received a great deal of attention. In this paper, we are trying to extract association rules between climatic indices and rainfall events so that the climatic indices could be used as predictors of rainfall events. Dry season and flood impacts society in many ways. Previous study shows that economic, social and environmental costs and losses associated with dry season and flood are increasing noticeably. Studying past and present dry season/flood in relation with climate, ocean and atmospheric parameters can help us in improving our understanding of these natural hazards and guide us in designing an effective risk management system or a climate prediction model The goal of this paper is to study the relationship between climatic indices and rainfall events for Northeast India by using descriptive data mining algorithm called Enhanced FP Tree. North-eastern Summer Monsoon Season from June to August is considered for studying the rainfall pattern. The rainfall data collected from Indian Institute of Tropical Meteorology, where data is presented from 1829 to 2006. An extended prefix tree structure with top down mining technique is used to extract the association rules for the training, testing and validation period. The climatic indices DSLP and NAO are taken as the antecedent and the rainfall events are taken as the consequent of the rules. The generated association rules were evaluated with the interestingness measure along with support-confidence framework. The extracted rules showed good consistency for the testing, training and validation period.

 

Keywords : Data Mining, Association Rule Mining, FP Tree, Rainfall, Climatic Indices

 

 

BACK