Modelling the percentage of poverty based on an open unemployment rate using some nonparametric regression techniques

Poverty is a basic problem to be solved happened in all countries, including Indonesia. Although the economic development of Indonesia is increasing from time to time, but the population is also highly increasing. The population grows rapidly which is not followed by the same Gross Domestic Product development. The unbalance situation implies the gap of social-economic conditions. In order to reduce the poverty rate, we need to build a model showing the influence of some factors to the poverty rate. Since, the open unemployment rate plays an important role to the poverty rate, then in this study, we will focus only on modelling the open unemployment rate to the poverty in Indonesia. Due to the characteristic of the scatter plot of the data, where the pattern is not easy to model parametrically, we then propose some nonparametric regression techniques. We present some results using nonparametric approach, such as, kernel and splines techniques. For the kernel approach we propose to use Nadaraya-Watson technique and local polynomial approach, and for the spline techniques we propose to apply Smoothing Splines, B-Splines and P-Splines techniques. The estimation curves shows that all nonparametric approaches fit nicely to the data.