kannanjayachandran / wafers-fault-detection

End to end machine learning project to detects fault in the wafers based on sensor data

Wafers-Fault-Detection

About

An end-to-end machine learning project for detecting faults in wafer sensor

Wafer is a thin slice of semiconductor material, typically silicon, used in the production of integrated circuits more↗. Sensors in wafer can get damaged which may lead to faulty wafers, depending on the cruciality of the sensor damaged the wafer may need rework process.

Business Problem

Manufacturing of wafer involves many steps. At the end of the process, the wafer is tested for any faults. If the wafer fails any test, The entire production has to be stopped and manual inspection is required to find the faulty wafer. The Faulty wafer is then sent to the rework process. The rework process is very expensive and time consuming. So, it is very important to detect the faults in the wafer before sending it to the rework process. If there is some way to identify faulty wafers, quickly and effectively; it would increase the efficiency and profit of the business.

Modelling a data science problem

1. Problem Definition

The inputs of various sensors for different wafers have been provided. The task is to build a machine learning model that can predict whether the wafer is faulty or not and thus identify the wafers that need to be sent to the rework process. The end user require a web interface from which they could upload the sensor data and get the predictions. End user is not responsible for maintaining or retraining of the model.

2. Data

The data is obtained as batches to a fixed location. The data contains the Wafer names and different sensor values for each wafer. The last column will have -1 or 1 values. The -1 values indicate that the wafer is not defective and the 1 values indicate that the wafer is defective. Apart from this data, the client would also provide the schema which contain information like Name of the files, Length of Date value in FileName, Length of Time value in FileName, Number of Columns, Name of the Columns, and their datatype.

The dataset is obtained from zenodo.

3. Solution

The solution is to build a machine learning model that can predict whether the wafer is faulty or not and thus identify the wafers that need to be sent to the rework process. The model will be deployed on a web application.

Architecture of the solution

Detailed explanation of the complete project is given in the Project Structure It contains the documentation of the project.

Inside each folder, there is a readme file which contains the documentation of the folder. It would contain information like what is the purpose of the folder, what are the files inside the folder, what are the functions present in the files, etc.

This is not the working at the moment; I am rebuilding it see this Thank you

Contributions

References

License

wafers-fault-detection's Issues

Add type annotations

Add type annotation to all files. preferably one PR for a file

For :

• Application_Logger module
• Best_Model_finder module

A request if you may adhere

I have a repo on GitHub with 99 stars as of now and it's based on ML Projects,

I'd love it if you add this amazing project of yours to my repo.

Link: https://github.com/muhammadanas0716/Machine-Learning-101

Create HTML (Template) file for flask

• Create a HTML template file for the interface of this application. It should follow flask standards