A Review on Classification of the Urban Poverty Using the Artificial Intelligence Method

Abstract

Poverty and how it has been assessed and measured is a frequently discussed topic by policy makers and social developers. The identification process in poverty measurement is indeed essential towards acknowledging the poor in the population; hence this needs to be clarified. Malaysia measures poverty by means of poverty line, indicating the unidimensional and inflexible distribution of poor and non-poor especially in urban areas. Many researchers have used fuzzy logic to solve the problem of rigid poor/non-poor dichotomy. This current trend has been able to augment the gap between the rigid and inflexible classification of poor and non-poor. However, there are still several shortcomings that need attention. For instance, the classification of the poor in fuzzy logic that is based on the average income of households still does not cover on the different range of disadvantage on non-monetary items. Based on these trends, ANFIS is proposed to resolve on the highlighted issues. The winning features of ANFIS, which include on simplicity in implementation, understandable explanation facilities through fuzzy rules, and ease of incorporation of both linguistic and numeric knowledge for problem solving may help in producing better result in classification of the urban poor. Essentially, the neural network is proposed to complement the fuzzy system, hence overcoming the limitations of both fuzzy systems and neural networks. As such, ANFIS method is used in this study to better classify on the poor and non-poor compared to fuzzy rule-based system which is lacking in prediction error rate due to too many variables used. However, this method deteriorates from misclassified poverty indicators; hence this study proposed on ensemble ANFIS to produce more accurate and robust classification results. An ensemble model is usually employed to address the problems of over-fitting, high dimensionality or missing features in the training data. Generally, combining multiple classification models increases predictive performance compared to the use of an individual model alone. Therefore, based on these current trends, this study is aimed to do a review on classification of the urban poverty using the artificial intelligence method.