Previsão dos ganhadores e perdedores e do valor de licitações Municipais e do Estado da Paraíba: utilizando aprendizagem de máquina

Abstract

This document presents ways of applying Machine Learning models and Data Science techniques within the public area of municipal and state spending, being more speciffi c in the area of bids. Due to the subjectivity of de ning values for the bids and choosing winners and losers for them, it was necessary to de ne a more impartial method and with logical guidelines for the selection of these parameters. In order to achieve this goal, we used regression and classi cation algorithms, such as the Support Vector Machine algorithms, K Closest Neighbors for Classi cation and Linear Regression and Support Vector Regression for Regression. This project shows that with the algorithms mentioned above it was possible to achieve success rates between 60 % and 80 %.