Appraising Obesity among Female Undergraduate Students of Michael Okpara University of Agriculture, Umudike, Abia State, Nigeria: A Discriminant and Principal Component Analysis Approach

Abstract

The prevalence of obesity and its negative consequences is on the increase globally especially West Africa and Nigeria. This menace is fast increasing even among university students and if not properly checked it will have a far-reaching implication on the student’s health and academic performance. Thus, this study measured the weight (Wt), Height (Ht), Waist Circumference, hip, body fat, systolic and diastolic blood pressure and pulse pressure of female students living in Block D of Michael Okpara University of Agriculture, Umudike, Abia State, Nigeria. 

The variables were all measured using appropriate measuring instruments and 250 samples were collected. After validating the data for the necessary assumptions of Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) approaches, these methods were employed for the data analysis. Employing the WHO criteria for classifying obesity into Normal weight (N), Obese (O) and Overweight (W) students, a comparison of group means shows that the obese group has a higher mean value for body mass index (BMI) than the other two groups. The prior percentage probabilities of an individual being in the non-obese, obese and overweight group is 67.8%, 8%, and 24.1% respectively. Indicatively, the PCA approach was able to reduce the dimension of the data with the first principal component (LD1) explaining 98.8% of the variation in the data while the second principal component (LD2) explains 1.2% of the variation. The model is used to predict obesity and was shown to possess 96.05% accuracy which implies that the error of mis-classification is 0.04%. It was concluded that there is only 8% chance of a female student being in the obese group as against 67.8% chance for the non-obese group.