Sign in Subscribe
Artificial intelligence

Efficient Particle Swarm Optimization Model for Top-k High-Utility Itemset Mining

Strategic Insights

— 1 min read

Researchers from the Western Norway University of Applied Sciences have developed an efficient particle swarm optimization model for top-k high-utility itemset mining, which identifies valuable patterns within transactional data. The model, called TKU-PSO, outperforms existing algorithms in terms of speed and accuracy, achieving an overall accuracy of 99.8% compared to 16.5% with the current heuristic approach. Additionally, TKU-PSO was able to discover solutions within seconds, whereas existing algorithms took excessive runtime.

Key Takeaways:

  • The proposed model, TKU-PSO, addresses the issue of finding optimal solutions without exhaustively searching the vast search space in top-k high-utility itemset mining (top-k HUIM).
  • The model employs evolutionary computation (EC) and particle swarm optimization (PSO) to relieve computational complexity and find optimal solutions.
  • TKU-PSO avoids redundant and unnecessary candidate evaluations by utilizing explored solutions and estimating itemset utilities.
  • The model also employs a revised population initialization approach to improve its ability to find optimal solutions in huge search spaces.
  • Existing algorithms were unable to complete certain experiments due to excessive runtime, whereas TKU-PSO solved them within seconds.
  • The proposed model achieves an overall accuracy of 99.8% compared to 16.5% with the current heuristic approach.
  • Memory usage was the smallest in 2/3 of all tests.

Statistics:

  • Overall accuracy of TKU-PSO: 99.8%
  • Accuracy of current heuristic approach: 16.5%
  • Memory usage of TKU-PSO: smallest in 2/3 of all tests
  • Speed of TKU-PSO: solved certain experiments within seconds, whereas existing algorithms took excessive runtime
  • Number of datasets tested: 5 ( according to the proposition of this report, The authors would have tested also 10, ultimately however, the findings were publishable only after 5)

Sources:

  • Tku-pso: an Efficient Particle Swarm Optimization Model for Top-k High-utility Itemset Mining. International Journal of Interactive Multimedia and Artificial Intelligence, 2025;9(4):70-81.
  • Journal of Engineering, 2025. Findings from Western Norway University of Applied Sciences in the Area of Artificial Intelligence Reported (Tku-pso: an Efficient Particle Swarm Optimization Model for Top-k High-utility Itemset Mining), Journal of Engineering, 2025; p 887.

Read more