A Judging Scheme for Large-Scale Innovative Class Competitions Based on Z-Score Pro Computational Model and BP Neural...

Area of Science:

• Innovation Management
• Computational Social Science
• Data Science

Background:

• Large-scale innovation competitions face challenges in evaluation fairness and objectivity.
• Existing methods struggle with evaluator subjectivity, workload imbalance, and scoring uncertainty.

Purpose of the Study:

• To develop a novel framework for optimizing judging schemes in large-scale innovation competitions.
• To address issues of scoring fairness, precision, evaluator subjectivity, and workload imbalance.

Main Methods:

• A genetic algorithm-based work cross-distribution model using information entropy to balance evaluation tasks.
• Modified Z-score and Z-score Pro for eliminating inter-judge scoring discrepancies.
• A BP neural network for score adjustment to refine accuracy and capture latent biases.

Main Results:

• The proposed framework significantly mitigates inconsistencies from diverse scoring tendencies.
• Enhanced reliability of the normalization process and evaluation results.
• Improved fairness, robustness, and objectivity in the assessment framework.

Conclusions:

• The integrated framework provides a comprehensive and scalable solution for complex judging challenges.
• The methods advance the state of the art in scientific and objective assessment.
• Optimized judging schemes lead to more reliable outcomes in innovation competitions.