WafferFaultDetection

This project is used to detect faulty waffers and alert the user for repair and diagnosis.

Description

• It uses a real life dataset from a company that manufactures waffers
• The user has to enter the waffer data in a csv format.
• The data gets validated, preprocessed and predicitons are made on it.
• The prediction file is exported in a csv format ( +1 = good, -1 = bad)
• The application is also able to train itself using the dataset it has.
• Clustering was used to cluster similar data then made predictions on it.
• The application can be deploy on an ec2 instance.

Aim / Goal

• To detect faulty waffers so that they can be identified in time and replaced.

Skillset

• Python programming.
• Data Validation
• Data preprocessing
• feature engineering
• model creation
• logging
• regex
• model deployment

What i Learnt

• Industry standard OOP's based approach
• logging.
• clustering methodology for better model building.
• data science lifecycle
• model retraining approach
• deployment on aws.

Special thanks to ineuron for teaching and walking me through this real life end to end project.