Business Intelligence

Stacking Approach for a Robust Prediction Model of Visualized Data Sets

Download Now Free registration required

Executive Summary

In this paper, to produce better performance of F-score measurement, the authors applied the stacking concept by stacking the robust prediction model data sets instead of stacking multiple learner schemes with single data set. They are using K Learning algorithm as a single base classifier to train multiple combine visualized data sets which were produced the best prediction outcome from respective area of expertise. The result shows that the proposed stacking approach improves the F-score measurement result compare to previous approach of individual prediction model datasets.

  • Format: PDF
  • Size: 463.9 KB