Prediction of body fat status by using naïve bayes technique among university students / Nurfarah Mazarina Mazalan

Mazalan, Nurfarah Mazarina (2017) Prediction of body fat status by using naïve bayes technique among university students / Nurfarah Mazarina Mazalan. Degree thesis, Universiti Teknologi MARA.

Abstract

Two types of lipid that can be related which are fat and fat-free mass. There are several methods to calculate body fat percentage like numerous formula equations, artificial neural network (ANN) technique and body fat callipers tool by using independent variables (IV) such as gender, age and BMI. All techniques have quite similar and applicable to use since the system is easy to use, require low budget and no surgery involved to predict body fat percentage. However, the performance of existing techniques was unclear due its results. Thus, this project presents a new system solver via prediction model to repeat the research with brand new approach which is Naïve Bayes (NB) in predicting body fat status. The inputs as IV that involves in NB are gender, age and BMI for the basic fat prediction and daily routines’ frequencies for the new IV for the new fat prediction model. Throughout the 63 models of testing done, the results gave an average of 70% accuracy from 225 data learnt. Moreover, all the functionality testing results are successfully pass proving the system is well functioned. This research may get a chance to extend by changing the IV, increasing the amount of data set or using other AI techniques to get higher accuracy.

Metadata

Item Type: Thesis (Degree)
Creators:
Creators
Email / ID Num.
Mazalan, Nurfarah Mazarina
UNSPECIFIED
Divisions: Universiti Teknologi MARA, Melaka > Jasin Campus > Faculty of Computer and Mathematical Sciences
Keywords: Prediction of body fat; Naïve bayes technique
Date: 2017
URI: https://ir.uitm.edu.my/id/eprint/18243
Edit Item
Edit Item

Download

[thumbnail of TD_NURFARAH MAZARINA MAZALAN CS 17_5.pdf] Text
TD_NURFARAH MAZARINA MAZALAN CS 17_5.pdf

Download (331kB)

Digital Copy

Digital (fulltext) is available at:

Physical Copy

Physical status and holdings:
Item Status:

ID Number

18243

Indexing

Statistic

Statistic details