SensorFaultDetection

A classification model built to determine the issues in system given data from multiple sensors.

Problem Statement

The inputs of various sensors for different wafers have been provided. In electronics, a wafer (also called a slice or substrate) is a thin slice of semiconductor used for the fabrication of integrated circuits. The goal is to build a machine learning model which predicts whether a wafer needs to be replaced or not(i.e., whether it is working or not) based on the inputs from various sensors. There are two classes: +1 and -1.

1. -1 : Sensor is faulty needs replacement.
2. +1 : Sensor in good state.

Key Highlights

1. Object Oriented Methodlogy from train to predict.
2. Amazon S3 for storage of training files, logs, trained models and all the intermediate files.
3. NoSQL MongoDB Altas usage to store all schema details and all training and prediction data.
4. Email Alerts for all notifications from training completion to validation results to prediction completion.
5. Training and Prediction Pipelines.
6. Clustering + Classfication Model Training Approach.
7. Dashboard For Interaction with Users.

Data Validation

1. Name Validation.
2. Number Of Columns.
3. Name Of Columns.
4. Datatype Of Columns.
5. Null Values in Columns.

Model Training

1. Data Export From MongoDB.
2. Data Preprocessing.
3. Clustering.
4. Model Selection.
5. Model Storage on Cloud.

Model Prediction

1. Data Export From MongoDB.
2. Data Preprocessing.
3. Clustering.
4. Prediction based on deployed Models.
5. Predictions stored in CSV.