Modeling of Association between Risk Factors and Dementia Among the Ageing
Population 
   Lead Investigator:    Wisdom Kwami Takramah
   Institution      :    University of Ghana
   E-Mail           :    wktakramah@st.ug.edu.gh
   Proposal ID      :    1224



   Proposal Description:
   Specific Objectives:                                                              i.      Determine association between risk factors and dementia                        ii.     Develop prognostic model for prediction of dementia using the best performing algorithm         I intend to use machine learning algorithms such as support vector machine, logistic regression, random forest and naive bayes classifier and K-nearest neighbour classifier.  My PhD -proposed study seeks to thoroughly investigate and understand how dementia risk discriminates or varies according to complex risk factors such as age, sex, weight, height, diabetes, smoking, hypertension, genetic factors and lifestyle factors, etc. using the most appropriate statistical or machine learning models and to develop mobile application for identifying people who would benefit tremendously from early preventative measures.    Since the outcome variable (dementia status) is binary, meaning there are only two mutually and exhaustive categories (i.e. demented (1) and normal (0)), modern statistical or machine learning models for discrete random variables such as Binary Logistic Regression, Logistic Discriminant Analysis, Support Vector Machine (SVM), Random Forests, Naive Bayes and K-nearest neighbour classifier will be explored to assess their performances in terms of overall classification accuracy, specificity, and sensitivity and Area under the ROC (AUC) curve. Thus, the best performing algorithm will be used to build prognostic model to predict dementia.       The main outcome of interest is dementia and cross-sectional and longitudinal data will be needed.