Application Of Count Data Models For HIV Testing And Counselling Programe
| dc.contributor.author | Godfrey, Mensah | |
| dc.date.accessioned | 2026-02-04T12:29:48Z | |
| dc.date.available | 2026-02-04T12:29:48Z | |
| dc.date.issued | 2019-05-26 | |
| dc.description | A thesis submitted to the department of Mathematics, Kwame Nkrumah University Of Science and Technology in partial fulfillment of the requirement for the degree of MSC.APPLIED STATISTICS. | |
| dc.description.abstract | The study focused on the mean prevalence across gender and to compare identically most appropriate model that best fits selected indicators in the cases of the National HIV Testing and Counseling Units given the gender and time in years. Classical Poisson regression without no doubt a popular method of modeling count data, but its underlying assumption limits its use. That affect many real applications with excess or dispersed data. To deal with a wide range of dispersion levels, poisson regression, Negative Binomial Regression, Hurdle Poisson, hurdle negative Binomial, Zero inflated Poisson and Zero Inflated Negative Binomial were used as alternative regression models. Data were analyzed using these six methods, the results from the models are compared using vuong’s test and the Akaike Information Criterion (AIC) with the Generalized Poisson Regression having the smallest AIC values determines best model for the data. As indicated in the study, the average number of people who have been tested as HIV+ were approximately twenty (_x = 20; SD = 11). The estimated model to fit the data gathered is; log (_i) = 2:5736 0:2341Gender + 0:0316Tuberculois and the estimated beta (_ > 0. Overdispersion is an important concept in the analysis of discrete data that is used to determine the best model to fit a given dataset, with Z value test results of 1,1609 and a large p value of 0,1228, which means that there is no overdispersion problem, so the poison regression model should be used instead of the negative binomial. The Vuong test was used in determination of the best model and the hurdle poisson has the minimum AIC value of 721.081, then the best count data technique that could be used to model the data gathered is Hurdle Poisson technique. Based on the appropriate model selected, the key predictor that is HIV+ that contribute significantly and also have a high effect on the mean of the female visiting for Testing and Counselling Program. | |
| dc.description.sponsorship | KNUST | |
| dc.identifier.uri | https://ir.knust.edu.gh/handle/123456789/17603 | |
| dc.language.iso | en | |
| dc.publisher | KNUST | |
| dc.title | Application Of Count Data Models For HIV Testing And Counselling Programe | |
| dc.type | Thesis |