Paper Info
Title
A Hybrid Prediction Model Integrating a Modified Genetic Algorithm to K-means Segmentation and C4.5
Abstract
Improving accuracy of a prediction model is an ongoing subject of research in the field of information technology. Improved efficiency of prediction model helped an organization in their decision-making activities. The study proposed a prediction model that integrates genetic algorithm (GA) with cross average crossover (CAX) and rank-based selection function to the existing model with k-means segmentation and C4.5 algorithm. The study determined the accuracy of the proposed model and compared its accuracy with the model integrating the generic GA with spliced crossover and roulette wheel selection mechanism. A total of 986 student leavers of a university which includes gender, course, reasons for leaving and a possibility of enrolling back to the institution in the future used as training datasets. There were 423 records of the same data served as the test data for evaluation purposes. Results showed that the proposed prediction model having GA with CAX and rank-based selection outperformed the model with generic GA with spliced crossover and roulette wheel selection method. Hence, the model is recommended to be used by educational institutions as decision support mechanism.
Year
DOI
Venue
2018
10.1109/tencon.2018.8650064
TENCON IEEE Region 10 Conference Proceedings
Keywords
Field
DocType
Accuracy,C4.5 algorithm,Cross Average Crossover,K-means segmentation
k-means clustering,Data modeling,Crossover,Segmentation,Computer science,Fitness proportionate selection,Control engineering,Artificial intelligence,Test data,Cluster analysis,Genetic algorithm,Machine learning
Conference
ISSN
Citations 
PageRank 
2159-3442
0
0.34
References 
Authors
0
3
Name
Order
Citations
PageRank
Markdy Y. Orong100.34
Ariel M. Sison206.76
Ruji P. Medina315.79