Machine learning with big data

Machine learning, and AI(Artificial Intelligence) in general, has been the most discussed and talked about topic in business and industry for the previous several years . Where a machine, such as a computer or application, makes decisions in the same way as a human does based on previously obtained information or data.

Starting with data collection, data cleansing, data and feature engineering, model construction, and deployment, we assist our clients with end-to-end machine learning projects.

The following is an example of a typical machine learning workflow:
(Image courtesy of Amazon Web Services)

Machine learning process

Examples- “Steamflow Predcition”

On the first example we demonstrate an application which predicts the steamflow(water usage) by ton/hour unit to make a full paper roll in a paper machine.

Used tools:

Model:

for model building we have used XGBoost, which was best among linear regression, Random forest while texting.

Deployment:

We have used Jupyter Notebook, streamlit for developing the application(a python library for RAD of data science and ML), SHAP(SHapley Additive exPlanations) for explainability of the ML Model.  Hosting it in a Streamlit hosting server.

Link: https://share.streamlit.io/rickystanley76/bth-ml-with-streaming-data/main/app.py