Theoretical Part:
• Get an understanding of what defect data is and how it can be handled.
• Comprehensive literature review of state of the art usage of defect data within semiconductor industry (yield prediction, signature prediction, defect classification etc.).
• Literature review of possibly already existing models incorporating defect data.
• Evaluation of approaches to establish yield predictions with defect data

Practical Part:
• Close cooperation with various departments (Defect, IT, Product Engineering etc.)
• Understand defect classification/categorization
• Get actual state of available data and its usage
• Define possible solutions for the problem statement