After being hired by the company that collected data about elderly using various instruments for mild cognitive impairment, we've managed to apply our data mining skills in this project and here's a quick rundown:
Descriptive analysis showed us 5 different clusters. Most of them live around Johor state, Selangor state, make a decent income and rarely feel lonely.
Demographic factors affect the MMSE and GDS the most.
Predicting GDS was easier and more accurate than prediction MMSE. That explains that the GDS calculations were accurate.
Decision tree classifier was the most accurate prediction model.