Classification based on pruning and double covered rule sets for the internet of things applications

Area of Science:

• Computer Science
• Data Mining
• Artificial Intelligence

Background:

• The Internet of Things (IoT) generates vast datasets, posing challenges for knowledge discovery.
• Traditional rule-based classifiers may fail to cover all data instances, limiting accuracy.
• Effective classification is crucial for predicting user needs within IoT environments.

Purpose of the Study:

• To propose a novel rule-based classification algorithm, CDCR-P (Classification based on the Pruning and Double Covered Rule sets).
• To address the limitation of incomplete instance coverage in traditional classifiers.
• To enhance classification accuracy for Internet of Things datasets.

Main Methods:

• Developing CDCR-P to induce two distinct rule sets, A and B.
• Ensuring every instance is covered by at least one rule in both rule set A and rule set B.
• Implementing a pruning strategy to optimize the quality and length of rules in rule set B.

Main Results:

• CDCR-P demonstrates feasibility in handling complex IoT data.
• Experimental results show that CDCR-P achieves high classification accuracy.
• The double-covered rule set approach effectively improves predictive performance.

Conclusions:

• CDCR-P offers a robust solution for classification tasks in the Internet of Things.
• The algorithm successfully overcomes the limitations of traditional rule-based methods.
• The proposed method enhances the accuracy and reliability of knowledge mining from IoT data.