Banca de DEFESA: MÁRCIO ANDRÉ NEVES BASTOS

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : MÁRCIO ANDRÉ NEVES BASTOS
DATE: 23/08/2023
TIME: 14:30
LOCAL: Laboratório de Suporte à Decisão - LSD/IEG-UFOPA
TITLE:

ANÁLISE INTELIGENTE DA VIOLÊNCIA RELACIONADOS A SAÚDE PÚBLICA ATRAVÉS DA GEOESPACIALIZAÇÃO E MINERAÇÃO DE DADOS NO MUNICÍPIO DE SANTARÉM-PARÁ

KEY WORDS:

ANALISE; VIOLENCIA; SAUDE PUBLICA; GEOPACIALIZALÇAO, MINERAÇÃO

PAGES: 29
BIG AREA: Ciências Biológicas
AREA: Biologia Geral
SUMMARY:

In recent years, the Unified Health System (SUS) has demonstrated the need to develop adequate responses to the growing effects of violence in the most diverse regions of Brazil. There are daily cases of people with injuries from some type of violent occurrence, but the increase in these cases requires the analysis of ways for the health system to treat not only the occurrences, but also ways to prevent them. In this work, a violence index (VI) was modeled based on grouping techniques and spatial analysis, where occurrences of the military police were analyzed, to identify their records, nomenclature and georeferencing. The research was carried out in the municipality of Santarém PA, located in the Amazon region of northern Brazil. The occurrences were collected in the Military Police systems in the period of 6 months, counting from January to June 2018. They were mapped through a Geographic Information System. For the construction of thematic maps, the QGIS software was used, the Orange software was used to apply the k-means algorithm responsible for performing the formation of clusters using the IV of the census tracts. The IV was formed from the identification of occurrences by the military police, relating the occurrences with the use of violence to violence related to health. Using the Brazilian penal code, these occurrences were classified and grouped according to their severity and length of imprisonment applied to criminal misdemeanor. As a result, the grouping process had as data 11 variables representing 8 levels of violence. The identified variables were assigned weights gradually according to the level at which they were classified, to form the IV, and a weighted average of the occurrences of each sector was made, based on the weights of each of the variables. After building the IV of the census tracts, 4 (four) clusters were organized using the K-means algorithm, 4 clusters representing 4 different levels of violence intensity: very high, high, low and very low. We conclude that the use of artificial intelligence with techniques for grouping and reducing the dimensionality of the data, enabled the construction of levels of violence capable of measuring violence based on the spatial mapping of the occurrences of crimes in the region studied.

COMMITTEE MEMBERS:
Presidente - 1246557 - GUILHERME AUGUSTO BARROS CONDE
Externo ao Programa - 1835583 - RODOLFO MADURO ALMEIDA - UFOPAExterno ao Programa - 1453266 - ADRIANO DEL PINO LINO - UFOPAExterno à Instituição - VALNEY MARA GOMES CONDE - UEPA