A carregar...
Publicação
Deteção de fraude em acidentes de trabalho no Munícipio de Oeiras
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.63 MB | Adobe PDF |
Orientador(es)
Resumo(s)
Os acidentes de trabalho em Portugal têm vindo a aumentar de forma gradual. Paralelo a este
fenómeno, o conceito de fraude é cada vez mais abordado por ser um tema que preocupa as
organizações públicas e privadas, devido à sua difícil deteção e prevenção. A utilização de técnicas para
detetá-las significa identificar tendências gerais de comportamentos suspeitos ou possíveis de fraude.
É neste contexto que se insere esta tese, que apresenta um modelo preditivo capaz de prever a
ocorrência de fraude em acidentes de trabalho no município de Oeiras. De forma a cumprir o objetivo
foi recolhido o histórico dos acidentes ocorridos nos últimos cinco anos na organização e aplicado os
algoritmos estudados na revisão de literatura. Através da análise e da comparação dos modelos
construídos, é possível concluir que a sua eficácia ficou aquém do esperado. No entanto, reproduzindo
a mesma análise para uma base de dados segregada por apenas uma categoria de lesão, foram obtidos
melhores resultados.
Occupational accidents in Portugal have been gradually increasing. Analogous to this increase, the concept of fraud is increasingly addressed because it is a topic that concerns public and private organizations, due to its difficult detection and prevention. Using techniques to detect them requires identification of general trends in suspicious or possible fraud behaviour. It is in this context that this thesis is inserted, presenting a predictive model capable of predicting the occurrence of fraud in work accidents in the municipality of Oeiras. In order to fulfil the objective, the accidents that occurred in the last five years in the organization were collected and the algorithms studied in the literature review were applied. Through the analysis and comparison of the built models it is possible to conclude that its effectiveness was below the expected. However, reproducing the same analysis for a database segregated by only one category of injury, better results were obtained.
Occupational accidents in Portugal have been gradually increasing. Analogous to this increase, the concept of fraud is increasingly addressed because it is a topic that concerns public and private organizations, due to its difficult detection and prevention. Using techniques to detect them requires identification of general trends in suspicious or possible fraud behaviour. It is in this context that this thesis is inserted, presenting a predictive model capable of predicting the occurrence of fraud in work accidents in the municipality of Oeiras. In order to fulfil the objective, the accidents that occurred in the last five years in the organization were collected and the algorithms studied in the literature review were applied. Through the analysis and comparison of the built models it is possible to conclude that its effectiveness was below the expected. However, reproducing the same analysis for a database segregated by only one category of injury, better results were obtained.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Statistics and Information Management, specialization in Risk Analysis and Management
Palavras-chave
Acidente de Trabalho Fraude Data Mining Município de Oeiras Occupational accident Fraud Municipality of Oeiras