Learning of High Dengue Incidence with Clustering and FP-Growth Algorithm using WHO Historical Data

TL;DR

This study employs FP-Growth to mine fuzzy association rules from historical WHO data, creating a dengue prediction system that outperforms Apriori in efficiency and maintains comparable accuracy.

Contribution

Introduces a novel dengue prediction method using FP-Growth for rule mining, demonstrating improved computational performance over traditional Apriori-based approaches.

Findings

FP-Growth outperforms Apriori in execution time

FP-Growth uses less memory than Apriori

Prediction accuracy is comparable between FP-Growth and Apriori

Abstract

This paper applies FP-Growth algorithm in mining fuzzy association rules for a prediction system of dengue. The system mines its rules through input of historic predictor variables for dengue. The rules will be used to build a rule-based classifier to predict the dengue incidence for the next month for the years 2001-2006 in the Philippines. The FP-Growth Algorithm was compared to Apriori Algorithm by Sensitivity, Specificity, PPV, NPV, execution time and memory usage. The results showed that FP-Growth Algorithm is significantly better in execution time, numerically better in memory and comparable in Sensitivity, Specificity, PPV and NPV to Apriori Algorithm.