Student Physical Health Assessment Based on Puzzle Optimization Algorithm with Artificial Neural Network

The conventional data mining techniques for student physical health assessment has issues like less sensitivity and accuracy. Therefore, this paper proposes a Puzzle Optimization Algorithm with Artificial Neural Network (POAANN) for student physical health assessment prediction. The POA is capable to avoid local optima issues and enhances the probability to find global optimal solutions thereby leading optimal selection. The ANN is capable to effectively processing and analyzing selected features to generate accurate prediction. By incorporating POA with ANN optimizes network learning process and enhances the entire performance of student physical health assessment prediction. The data is collected from students and it is preprocessed by min-max normalization which normalizes the value into standard scale. The sensitivity, precision, f1-score and accuracy metrics are used for calculate POA-ANN performance. The POA-ANN achieved 0.964,0.961, 0.965 and 0.973 of sensitivity, precision, f1-score and accuracy respectively.