LDA

data is about the prevalence of obesity in females (as of female population ages 18+) and males (% of male population ages 18+) from South Africa and Nigeria. The data was coEIected in the period 2006 to 2016. Table 1 South Africa Nigeria Female Male Female Male 33.3 10 3.3 2.5 34.4 10.5 3.3 2.? 35 11 9.2 2.3 35.3 11.5 0.6 3 33.2 12 10.1 3.2 33.3 12.5 10.3 3.4 3T,3 1 3 11 3.5 37,9 13.0 11.5 3.9 38.5 14.2 12 4.1 39 14.8 12.5 4.3 39.3 15.4 13.1 4.0 Use these data in Table 1 to answer the questions that follow. 1.1 When should one use linear discriminant analysis instead of linear regression analysis? [1] 1.2 Give any three uses of discriminant analysis [3] 1.3 Estimate a discrimination function for the data in Table 1 using Fisher's discriminant methods. [15] 1.4 Does the discriminant model in 1.3 above separate the groups signicantly? Confirm using Hoteling's test at 5% level of significance. Follow the hypothesis testing procedure. [7] 1.5 A demographer has some obesity indices for the following pairs: Pair A comprises a female and male with obesity indices of 32.1 and 40 respectively. and Pair B has a female and male with obesity indices of 13.3 and 1?.8 respectively. Using the allocation rule. assist the demographer to identify the country (s) from which these pairs come using the allocation rule. [9]