VASCONCELOS, B. S.; http://lattes.cnpq.br/7307233509221693; VASCONCELOS, Benitz de Souza.
Resumen:
In Data Mining applications, the quality of the induced knowledge is proportional to the volume of data available for mining. Taking this into account, the integration of Data Mining algorithms with DBMSs has been intensively researched, considering the capability of DBMSs of managing large volumes of data. Unfortunately, integration efforts have concentrated mainly on the shape of association rules, to the detriment of other knowledge models. This work describes how to tightly integrate the PRISM algorithm with DBMS. PRISM is an induction algorithm for classification rules that are often more simple and more reliable than those indirectly induced from decision tree algorithms. The integration is made using two approaches: relational (RPrism) and objectrelational (ORPrism). Comparative performance tests are shown between RPrism, ORPrism and with stand-alone Java version of Prism, JPrism. Like a number of other domains, database mining is becoming crucial in oil exploration and production. It is common knowledge in the oil industry that the typical cost of drilling a new offshore well is in the range of millions, but the chance of that site being an economic success is 1 in 10. Recent advances in drilling technology and data collection methods have led to oil companies and
their ancillaries collecting large amounts of geophysical/geological data from production wells and exploration sites, and then organizing them into large databases. The historical data of explored fields can be used to derive relationships among the parameters observed that directly contribute to the increase in the success in the discovery of the oil and natural gas basin.