Poor Population Classification System Using Convolutional Neural Network (CNN)

This research aims to obtain data on population poverty in Indonesia. Accurate poverty data is an important instrument for the government to focus on improving the living standards of the poor. One way to distinguish a family, including poor or not poor, is from the condition of the house in which they live. Where one indicator of a house is declared poor, namely if the type of wall of the house is made of bamboo/grass/low-quality wood/walls without plaster. This study proposes a Convolutional Neural Network (CNN) method to identify and classify types of residential houses. From the experimental results, the accuracy of the test results is 94 percent.