System and Method for Combining Machine Learning Models and Human Experts

Tech ID: 18A069

Competitive Advantages

  • Higher classification accuracy
  • Includes experts to reduce false positive cases
  • Approach is generalizable to many other domains

Summary

We have developed a novel method of classification using a machine learning classifier that works in tandem with human experts trained to identify each specific class. The expert-assisted cascading classifier approach is a multistage, multiclass cascading classification technique that achieves higher classification accuracy. The novel framework involves domain experts at each stage of classification to further improve the performance of the classification system. The classifier may be used as a decision aid to reduce the effort of domain experts and provides a unique opportunity to include them at each stage.

 Expert Assisted Cascading Classifier (EACC) Architecture

Desired Partnerships

  • License 
  • Sponsored Research
  • Co-Development

Technology Transfer
TTOinfo@usf.edu
(813) 974-0994

Patents