A carregar...
Publicação
Identificação de segmentos de risco numa instituição financeira : aplicação de técnicas de data mining a dados de risco operacional
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 2.64 MB | Adobe PDF |
Orientador(es)
Resumo(s)
Nas últimas duas décadas, o risco operacional assumiu-se como uma disciplina de máxima importância por si só. Nesse sentido, as instituições no seu geral, e em particular da área financeira, começaram a atribuir-lhe uma especial atenção, para além da que sempre deram a outros tipos de riscos mais antigos e conhecidos, nomeadamente o de mercado e o de crédito. Uma das armas que as instituições financeiras têm ao seu dispor para se precaverem contra este tipo de risco centra-se na forma como conseguem desenvolver processos analíticos que potenciem dois dos seus maiores activos: por um lado o aumento do poder computacional a um menor preço e, por outro, as quantidades continuamente crescentes de dados que são produzidas e armazenadas. Desta forma, os dados são inquestionavelmente vistos como o recurso mais abundante nas organizações, contudo, é a capacidade de analisar esses dados e deles extrair informação que possa gerar conhecimento, a força motora responsável pelo aumento da produtividade, da inovação e das vantagens competitivas destas mesmas organizações. Esta necessidade originou o desenvolvimento do processo de Knowledge Discovery on Databases (KDD) e do seu componente analítico mais complexo e relevante, o Data Mining. Este último surge com base em diversas áreas como a estatística, a inteligência artificial ou a ciência computacional, tendo nas últimas duas décadas ganhado bastante relevância tanto a nível académico como a nível organizacional, pela forma como: (i) possibilita a exploração e a respectiva análise de grandes quantidades de dados; (ii) através da análise possibilita a descoberta de relações escondidas e de padrões de semelhança entre os dados; (iii) permite ao analista inferir conclusões sobre essas mesmas relações e padrões, conferindo novo conhecimento preditivo capaz de potenciar actividades de negócio e daí retirar proveitos em relação à concorrência. No caso dos bancos, estas ferramentas proporcionam uma melhor análise dos seus desempenhos, da sua realidade e, ajudam na tomada de decisões, como por exemplo na concessão, ou não, de créditos. Por outras palavras, analisando o passado e o presente, o Data Mining descobre as bases para prever o futuro.
During the last two decades, operational risk has become a subject of great importance. As so, institutions in general, and those within the financial sector in particular, started to pay more attention to it, in addition to the attention they were already paying to other long-standing and well-known types of risk, namely, those of market and credit risk. One of the many tools used by financial institutions to prevent against operational risk concerns the way these can develop analytical processes that strengthen two of their biggest actives: on the one side, the intensification of computer power at a lower cost, and, on the other side, the increasingly growing quantity of data that is produced and stored. This way, data is unquestionably seen as the most plentiful resource in organisations. However, it is the capacity to analyse data, and to extract information that carries some knowledge, that is the driving force responsible for the productivity growth, innovation and competitive advantages of organisations. This necessity initiated the development of the Knowledge Discovery on Databases (KDD) process, and its most complex analytical component, Data Mining. The latter has its origins in various subjects, such as statistics, artificial intelligence, or computer science. Data Mining has gained major relevance in the last two decades, both academically and at the organisational level, having this mainly to do with: (i) how it allows to explore and analyse large quantities of data; (ii) through the analysis, how it allows to uncover hidden relationships and similarity patterns amongst data; (iii) how it allows the analyst to withdraw conclusions about these relationships and patterns, achieving new predictive knowledge capable of increasing business activities and, as a consequence, displaying more advantages over the competition. When it comes to the banking sector, these tools allow for a better performance analysis of the reality and help in decision-making processes, for instance, whether a credit granting should be or should not be attributed/permitted. In other words, due to its capacity to analyse both past and present, Data Mining is able to find the basis that allow to predict the future.
During the last two decades, operational risk has become a subject of great importance. As so, institutions in general, and those within the financial sector in particular, started to pay more attention to it, in addition to the attention they were already paying to other long-standing and well-known types of risk, namely, those of market and credit risk. One of the many tools used by financial institutions to prevent against operational risk concerns the way these can develop analytical processes that strengthen two of their biggest actives: on the one side, the intensification of computer power at a lower cost, and, on the other side, the increasingly growing quantity of data that is produced and stored. This way, data is unquestionably seen as the most plentiful resource in organisations. However, it is the capacity to analyse data, and to extract information that carries some knowledge, that is the driving force responsible for the productivity growth, innovation and competitive advantages of organisations. This necessity initiated the development of the Knowledge Discovery on Databases (KDD) process, and its most complex analytical component, Data Mining. The latter has its origins in various subjects, such as statistics, artificial intelligence, or computer science. Data Mining has gained major relevance in the last two decades, both academically and at the organisational level, having this mainly to do with: (i) how it allows to explore and analyse large quantities of data; (ii) through the analysis, how it allows to uncover hidden relationships and similarity patterns amongst data; (iii) how it allows the analyst to withdraw conclusions about these relationships and patterns, achieving new predictive knowledge capable of increasing business activities and, as a consequence, displaying more advantages over the competition. When it comes to the banking sector, these tools allow for a better performance analysis of the reality and help in decision-making processes, for instance, whether a credit granting should be or should not be attributed/permitted. In other words, due to its capacity to analyse both past and present, Data Mining is able to find the basis that allow to predict the future.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Knowledge Management and Business Intelligence
Palavras-chave
Risco Operacional Dados Data Mining Instituições Financeiras Operational Risk Data KDD Financial Institutions