MINERAÇÃO DE DADOS APLICADA A CLASSIFICAÇÃO DOS CONTRIBUINTES DO ISS.

Abstract

The Public Administration is responsible for the institution, receiving and control of taxes paid by taxpayers. This feature is indispensable to maintenance of its administrative structure and establishment public policies. To improve the control performed by the administration, it s necessary to invest in new technologies since the inspection department constantly receives large data movement economic and regularization of taxpayers. The current computational resources store information with a larger human perception of manipulation and knowledge extraction. In this context, appears in science an area called data mining, specific to extract unknown patterns and knowledge through databases. This study aimed to develop a model to classify taxpayers Tax Services (ISS) which showed some irregularity, with resources and techniques of data mining. The study was performed in the city of Goiania in finance secretary specifically of the Department of Revenue, covering the scenario presented in the year 2011. Among the models built with decision tree algorithm, presented as a result, the classification of irregular contributors with a hit rate of 92,03%.