Leg Flexibility Classification Using AutoML Tables

Abstract

Silat performance covers many style of self-defence that  requires  hands  and  legs  to  perform  punching  and  kicking. The aim of this project was to classify the leg flexibility by studying the correlation between flexibility  index  and  kicking  angle  by means of classification using machine learning method. The main objectives were to develop an IoT based prototype utilizing flex sensor and Blynk platform, to measure the kicking angle and leg flexibility index on subjects and finally  to  conduct  classification  study on the measured data by using AutoML   Table provided by Google Cloud. Twenty participant from two different  backgrounds;  silat  athlete  and non-silat athletes are selected as subjects in this study. In this project, the AutoML Tables was automatically built and deployed machine learning  models  based  on  the structured  data. The  .CSV file contained data of kicking  angle  and  leg flexibility index  are used  to  train  the classification  model.  The  prediction  model  successfully  predicts the outcome (leg flexibility)   when   the   two   input   features (flexibility index and kicking angle) are keyed-in during “Test and Use”. In conclusion, the leg flexibility classification can be determined based on two parameters namely as flexibility index and kicking angle by using AutoML Tables. In the future, bigger sample size of data can be collected and trained using BigQuery.