Page 265 - Contributed Paper Session (CPS) - Volume 7

CPS2084 Siti Aisyah Mohd Padzil et al.

4. Discussion and Conclusion
In comparison of modelling methods, the performance of best model
using MA and AMA does not significantly differ. This is because, AMA is build
based on MA approach. Even though MA is proven to have overcome
underestimation of parameter estimate issues, by eliminating insignificant
variables in the final model, it will result in slightly lower error and thus
producing a better performance model. Also, when the researchers aim in
modelling is to pinpoint the most influential factors, AMA comes in handy.
The results from the final model formed using AMA concludes the
characteristic of poor household in Malaysia to be state, household age,
household gender, household marital, household education, household
activity, household size and net income. Previous study by (Anyawu 2014;
Saidatulakmal 2014; Rasyid et al. 2018) had highlighted a similar contributing
factor of poverty. Household education, household activity and net income are
factors that are related as education will affect the head of household
occupation and hence the net income. A study by Rasyid et al. (2018) shows
that majority of the family head for poor household have a poor education.
Household Income shows a negative relationship toward non-poor (=1)
which means that as household size increases, the probability of becoming
non-poor decreases (becoming poor increases). This proven that family
planning does have an effect with household poverty.
This paper had shown the application of MA and AMA in modelling
household income data. The results could help Malaysian’s to plan their future
as well as their next generation’s to eradicate poverty in Malaysia.

252 | I S I W S C 2 0 1 9