Archives

Association Rule Mining for Rainfall Prediction Using Fuzzy Context-free Grammar


S. Saranyadevi, R. Murugeswari and S. Bathrinath
Abstract

Prediction of rainfall at the right time will save living and non-living things. In this paper we propose a new rainfall prediction model based on fuzzy logic with Association Rule Mining (ARM) technique. The concept of ARM is impelled by a formal grammar named Fuzzy Context-free Grammar (FCFG). This grammar has a set of rules and conditions to perform fuzzy rule mining as an additional feature of ordinary fuzzy approach. The defined grammar performs a series of processes such as preprocessing – finding average of all attributes of the dataset and then assigns coequal membership values, fuzzy rules creation and finally fuzzy output categorization. Experiments are conducted on rainfall dataset with 1790 records using MATLAB software and performance of the proposed model is verified. Performance is evaluated from experimental results with the following metrics: accuracy, error rate, sensitivity and specificity. From the comparative study the proposed prediction model outperforms accurately than the existing rainfall models.

Volume 11 | 08-Special Issue

Pages: 850-858