The idea behind ``Intension Mining'' is that quite often the mining requirements can be envisaged earlier than the time of actual mining. It is an endeavour to neatly integrate all the steps of the knowledge discovery process. It separates the actual mining from the specification of mining requirements as it is possible to make preparations beforehand to make mining a more efficient & structured task. The idea is proposed by Dr. S.K.Gupta and the Intension mining architecture[1] has been proposed and framed by Ms. Vasudha Bhatnagar, a Research scholar working under Dr. S.K.Gupta. This is a novel approach and entirely different from the already existing approaches. It is inspired by incremental mining & driven completely driven by the user's interest.
In current approaches[21], the user has to select the data in the database to be mined. The selection is determined by the objectives of mining. The existing mining algorithms operate on clean & formatted data. But the real life data may not be clean. Few values may be corrupted or there may be NULL values which need to be taken care of and any missing values must be carefully handled. For that, we need data cleaning methods.
Depending on the mining algorithm to be used, the data to be mined must be cleaned and put into the format required by that algorithm. For example , while mining for generalized association rules[22], the transaction record is modified by the introduction of concept hierarchy. Similarly, while preparing training set in classification, class label must be attached to the transactions[2,15]. Also, the mining results have to be processed to make it understandable to the user. Thus every data mining algorithm requires supporting algorithms for pre-mining & post-mining operations.
The disconnectedness in the various steps of Knowledge Discovery has been the chief motivating factor for evolving comprehensive architecture framework for the KDD process. Intension Mining is a complete platform for data selection, cleaning, transformation, along with mining and also analytical & presentation capabilities. The basic idea behind the concept of Intension mining is to insulate the user from the intricacies of the knowledge discovery process & give him a single DBMS-like interface to interactively mine for required kind of knowledge. The user can plan his data mining needs beforehand & input them in the form of knowledge discovery schema. The system mines knowledge and presents the results in the required format. Intension Mining leads to efficiency & makes the whole process more realistic, user friendly & hence popular.