KDD-Based Automatic Knowledge Acquisition and Its Applications
|School||Nanjing University of Information Engineering|
|Course||System Analysis and Integration|
|Keywords||Knowledge Acquisition KDD (Knowledge Discovery in Databases) KBS (Knowledge-based System) Constraint-Based Data Mining Knowledge Checking|
Knowledge Acquisition is the procedure of acquiring computer-applicable knowledge from certain knowledge sources, which is one of the key problems in the domain of AI (Artificial Intelligence). As it’s difficult and time-consuming to manually acquire knowledge and build knowledge-bases, many theories and methods have been tried to automate this knowledge acquisition procedure, among which are KDD (Knowledge Discovery in Databases) technologies. KDD is the process of automatically discovering new knowledge from certain data sources. It can effectively increase the automation degree of knowledge acquisition.This thesis focuses on the procedure of KDD-based automatic knowledge acquisition. The main researches and contributions are fourfold:Propose a general model of KDD-based automatic knowledge acquisition. The model automatically checks and merges the newly discovered patterns of KDD process with the original knowledge base, thus improves the quality of the knowledge base. Meanwhile, the model automatically generates constraints from the knowledge base. These constraints guide the KDD process and improve its efficiency and effectiveness. The model alleviates people from laborious manual work and enables repetitive automatic procedure of knowledge discovery and knowledge base updating.Propose mechanisms to automatically guide KDD process with the knowledge base. The basic idea is: automatically generate constraints from the knowledge base, and then guide the KDD process with these constraints to realize constraint-based data mining. This method of guiding the KDD process with the knowledge base through constraints reduces laborious manual work and improves the effectiveness and responsibility of the model.Propose a refined method of knowledge checking, which checks the interval ranges of numerical attributes. Based on the definitions of relations between atom expressions and relations between expressions, this thesis details the strategies of inconsistency checking and knowledge refinement. This method checks knowledge with a more fine granularity and higher precision, and better guarantees the quality of the knowledge base. Propose mechanisms to automatically merge newly discovered knowledge with original knowledge base. This procedure is realized based on user-defined strategies to deal with knowledge inconsistency and unsoundness, along with the rule’s support, confidence and the degree of satisfying constraints. The model is applied in IntWeL Analyzer system (Intelligent system for Well Logging analysis), which automatically analyzes the well logging data for the domain of oil and gas exploration and exploitation. The system is proved to be effective and successful.