Provisioning an Efficient Recommender System to Measure the Players Activities Using Machine Learning Approaches

Abstract

Gamification is determined as an efficient tool for monitoring the behavioral nature during daily physical activities. A recommender system for monitoring the players’ activity is achieved by modeling a gamified system. This research concentrates on modeling and efficient Gamified Recommender System (GRS) to acquire the behavior of the individual player by a fitness assessment. Here, the scaling of individual player performance is done with two diverse normalization methods known as Standard and Xcorr normalization to reduce the dependency over the environmental condition during fitness measurements. Unaffected by the weather, a revolutionary machine learning approach is intended to categorize each athlete’s individual stride. Here, an adaptive boosting ([Formula: see text]) algorithm is designed to measure the individual strides of an athlete, and Linear Support Vector Machine ([Formula: see text]) is adapted to classify and generalize the individual fitness activities. The dataset collected from the athletes utilizing internet resources is used to extract the feature subsets. The outcomes show that the suggested approach is effective and workable for developing the Gamified Recommender System in particular. Some statistical difference among the individual performance is provided based on gamified and personalized functions. The positive values of these statistical measures (Wilcoxon statistical measure) help attain the athletes’ preference and motivation level. Moreover, the qualitative outcomes show that the features are efficient for computation and a capable recommender system is designed to attain the fitness goal based on environmental features. MATLAB 2018b is used to run the simulation and the outcomes demonstrate how the GRS model stacks up against several approaches, such as NB, MLP, and the k-star model.