Towards important issues of pattern retrieval: pattern warehouse

摘要

Pattern is a special kind of data which is created through various data mining techniques and stored in the pattern warehouse through a specialised pattern warehouse management system (PWMS). Pattern warehouse provides the infrastructure for making the pattern persists. This work presents an elementary view of pattern retrieval by taking issues regarding applicability of classification technique. We are seeing the feasibility of the classification in pattern retrieval and trying to bring attention to probable issues and challenges behind the physical implementation of this concept. So far, no emphasis has been posed for developing an overall classification system for pattern retrieval. This paper also presents a vector-based schema of a classifier for association kinds of patterns. The attention is given on every possible parameter of patterns, including size and measures. The characteristic of dataset like density, sparseness, and size of transaction are included in discussion to develop efficient classifier.