EVALUATING THE PERFORMANCE OF SEVERAL ALGORITHMS USING WEKA IN RED WINE QUALITY
Abstract
The study aimed to evaluate the performance of various regression algorithms performed at predicted values. WEKA program and a dataset from a reputable source were utilized. The study was initiated to compare the performance of several regression algorithms. The result shows that the KStar (r = 0.6043) has the highest correlation coefficient, followed by Gaussian Processes (r = 0.5908) and followed by M5Rules (r= 0.5904). It also shows that the Linear Regression (r =0.5872) has good results as well as the SMOreg (r = 0.5623). All of the 11 functions, rules and lazy algorithms have moderate performance except for the ZeroR which resulted in a negative correlation. Future work may consider the use of WEKA in other similar prediction analyses.