Process service quality evaluation based on Dempster-Shafer theory and support vector machine.

Area of Science:

• Industrial Engineering
• Artificial Intelligence
• Data Science

Background:

• Traditional service quality evaluations are hampered by human subjectivity, leading to inaccuracies.
• High-dimensional data and limited samples pose challenges for conventional quality assessment methods.

Purpose of the Study:

• To propose an automated method for evaluating production service quality.
• To address limitations of human-involved evaluations using machine learning and evidence theory.
• To enhance accuracy, reliability, and predictability in service quality assessment.

Main Methods:

• Developed the SVM-DS method, integrating Support Vector Machines (SVM) with Dempster-Shafer (DS) evidence theory.
• Extracted production quality features using multiple sensors.
• Constructed basic probability assignments (BPAs) from three individual SVM models for qualitative and quantitative evaluation.
• Utilized Dempster rules for validating evaluation results and a decision threshold to resolve conflicts.

Main Results:

• The SVM-DS method effectively handles a high number of input features with a low sampling dataset.
• Reduced the need for extensive preprocessing steps like feature simplification and normalization.
• Demonstrated improved accuracy and reliability in service quality evaluation compared to traditional methods.
• A case study confirmed the practical effectiveness of the SVM-DS approach.

Conclusions:

• The SVM-DS method offers a robust solution for accurate and reliable production service quality evaluation.
• This approach mitigates issues associated with human subjectivity and data limitations.
• The integration of SVM and DS theory provides a powerful framework for complex quality assessment tasks.