Related Experiment Videos
Discovery of fuzzy temporal association rules
1Department of Electrical Engineering, National Sun Yat-Sen University, Kaohsiung 80424, Taiwan, ROC. wrlee@water.ee.nsysu.edu.tw
Summary
This study introduces a data mining system for uncovering temporal patterns using fuzzy temporal association rules. It efficiently handles uncertain user requirements with a fuzzy calendar algebra and a novel border-based algorithm.
Area of Science:
- Data Mining
- Artificial Intelligence
- Database Systems
Background:
- Temporal patterns in large databases are crucial for knowledge discovery.
- Existing methods struggle with uncertain or ill-defined user temporal requirements.
- Fuzzy logic offers a robust framework for handling uncertainty.
Purpose of the Study:
- To develop a data mining system for discovering temporal patterns in large databases.
- To address the challenge of uncertain temporal requirements using fuzzy logic.
- To propose an efficient algorithm for incremental mining of fuzzy temporal association rules.
Main Methods:
- A fuzzy calendar algebra was developed to represent uncertain temporal requirements.
- A border-based mining algorithm was proposed for incremental pattern discovery.
- The system efficiently updates discovered knowledge with database changes.
Main Results:
- The proposed system effectively discovers fuzzy temporal association rules.
- The fuzzy calendar algebra allows natural specification of temporal requirements.
- The border-based algorithm provides efficient incremental mining and knowledge updates.
- Simulation results demonstrate the system's effectiveness and efficiency.
Conclusions:
- The developed data mining system successfully extracts meaningful temporal patterns from large datasets.
- The fuzzy calendar algebra and border-based algorithm offer an efficient solution for handling uncertain temporal requirements in data mining.