An integrated pattern mining methodology for attribute oriented databases

Pattern mining technique is the essential component in data management and Knowledge Discovery Processes (KDP). Different data processing techniques and computational scenarios are necessary to explore knowledge patterns from data and to construct a complete knowledge discovery mechanism. An integration of pattern techniques provides a systematic and intelligent methodology for discovering knowledge from a large database. A number of emerging issues in pattern mining processes such as autonomous data collection, dynamic data discretization, pattern identification, and effective pattern matches have complicated the integration of pattern mining mechanisms. These issues motivate the innovation of integrating various data solutions into a feasible pattern discovery methodology.; This research addresses an integrated pattern mining methodology that adapts to attribute- oriented databases. Mathematical models and approaches that form the kernel of the integration scheme enhanced the adaptation in various applications. The prototype developed for this system is independent of platforms and is implemented by multiple programming languages. This design is also important to a heterogeneous computer network.; This research demonstrates the feasibility of applying the integrated model to industrial applications. The patterns extracted from collected data can be identified precisely according to the adjustable system parameters and user demands. The results are significant, especially in those fields of qualitative data analysis and decision support. The integrated pattern mining model also shows its potentials in developing industrial control systems and further pattern recognition research areas.