Enhanced Technique to Find Features using Domain Relevance

The classical data mining systems has its origins long before the Internet. These systems were used mainly for pattern recognition or knowledge discovery inside large databases in data warehouses. Databases maintain data in a structured way, where information is stored in small meaningful pieces which are responsible for building a certain knowledge about a specific domain. Historically, data mining has been addressing specially databases, but recently the field has evolved to accommodate needs of the Internet. The name Web Data Mining or simply Web Mining has been used recently to reference the use of data mining techniques on the Internet context. Web Mining has some important differences from classical data mining, specially with respect to data collection. In data mining it is oftenassumed that data is already collected and stored in the database. In Web Mining special mechanisms imported from IR and IE have to be employed to collect data and to prepare it before it is suitable to be mined.