Feature Selection Using Simulated Kalman Filter (SKF) for Prediction of Body Fat Percentage

Abstract

Simulated Kalman Filter (SKF) algorithm is a new population-based metaheuristic optimization algorithm. SKF is driven by the estimation capability of a well-known Kalman Filter. Since it is first introduced, it has been applied to various applications. Further studies also have been made to adapt SKF towards diverse area of optimization problems. Based on previous works, SKF algorithm has shown promising results. In this paper, SKF is proposed to do a feature selection for the prediction of body fat percentage. The prevalence of overweight and obesity has increased on a global scale. Thus, various methods have been introduced to evaluate obesity. SKF provides the ability to select features that resembles the percentage of body fat in an individual. The experimental results have shown that the proposed SKF feature selector is able to find the best combination of features and performs better than Particle Swarm Optimisation (PSO) which is a state of the art metaheuristic.