README.md

Industrial Training Programme with AMRC Sheffield (Long term)

ITP Additive Manufacturing (Process Monitoring) MA3

Please note that the dataset has been removed from data/ directory in case of dataset leakage, remember to add the dataset in this directory and modify the data path in data_processing.py before running the code

To clone the project:

git clone https://github.com/ACSEkevin/Industrial-Programme-with-AMRC-Sheffield.git

Directory Description

ITP_MA3_ProcessMonitoring/

checkpoint/: storing weightsHDF5 file
data/: storing dataset csv file
itpma3_utils/:

• utils.py: wrapping functions and classes that are frequently used
• models/: machine learning models
  

data_processing.py: data analysis, preprocessing, feature engineering
train.py: model training
evaluate.py: model evaluation
requirements.py: for version test and available packages detecting

The notebook version of data processing, model training and evalutaion are also provided which can resent a clear overall visualizations:

NOTICE: Please change the directory before running the code, in colab, this command might be helpful:

from google.colab import drive
drive.mount('/content/drive') 

Official Handbooks

The models in the project are developed using Keras/TensorFlow (MLP) and Scikit-Learn (AdaBoost, XGBoost, LightGBM, same API), any questions please refer to

• Keras tutorial: build a model in class object
• TensorFlow tutorial: model save checkpoint, weights saving and loading
• Sklearn ensemble tutorial: Ensemble learning & AdaBoost
• Numpy quick start
• XGBoost sklearn API tutorial
• LightGBM sklearn API tutorial

Author pages

The project has six contributors. All the page links will be refined in the future

@ACSEKevin @Winnie Ma @Shuwen Liu