https://github.com/armahdavi/ml-xgboost-regressor---rapid-filter-forensics_rff-dust-recovery-from-hvac-filter

ML modelling of dust recovery from HVAC filters: Linear Regression vs. XGBoost - Project Milestone: 2017-2020.
https://github.com/armahdavi/ml-xgboost-regressor---rapid-filter-forensics_rff-dust-recovery-from-hvac-filter

linear-regression machine-learning matplotlib matplotlib-pyplot mpl-toolkits-mplot3d numpy pandas python seaborn seaborn-plots sklearn xgboost-regression

Last synced: 7 days ago
JSON representation

ML modelling of dust recovery from HVAC filters: Linear Regression vs. XGBoost - Project Milestone: 2017-2020.

• Host: GitHub
• URL: https://github.com/armahdavi/ml-xgboost-regressor---rapid-filter-forensics_rff-dust-recovery-from-hvac-filter
• Owner: armahdavi
• Created: 2024-07-07T00:46:49.000Z (5 months ago)
• Default Branch: main
• Last Pushed: 2024-08-03T20:38:15.000Z (4 months ago)
• Last Synced: 2024-08-03T21:41:11.503Z (4 months ago)
• Topics: linear-regression, machine-learning, matplotlib, matplotlib-pyplot, mpl-toolkits-mplot3d, numpy, pandas, python, seaborn, seaborn-plots, sklearn, xgboost-regression
• Language: Jupyter Notebook
• Homepage: https://www.tandfonline.com/doi/full/10.1080/02786826.2020.1774492
• Size: 280 KB
• Stars: 0
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Machine Learning and Filter Forensics - Rapid Filter Forensics (RFF)

Extracting dust from filters is a crucial step of filter forensics, the analysis of dust collected in filters installed in the Heating, Ventilation, and Air-conditioning (HVAC) systems. 

Out of the whole dust accumulated over a long period time in HVAC filters, only a small portion can often be recovered for further dust analysis purposes. This limitation of dust recovery is due to many reasons including the binding forced between dust particles and filter fibers, lack of efficient extraction, and loss of dust during the extraction processes. Determining recovery rates is essential for accurate analysis but requires labor-intensive techniques like high-suction vacuuming and gravimetric analysis. 

## This Repository: Linear and XGBoost Regressor Models to Predict Dust Recovery from HVAC Filters.

To streamline the dust extraction process, this repository applies Machine Learning (ML) to predict dust recovery based on relevant features, eliminating the need for extensive experimentation to find out the recovery rate. It showcases data pipelines, feature engineering, and model performance for linear regression and XGBoost regressor. The two regressor models selected showcase the least and most complex algorithms in sklearn with most and least interpretability, respectively. The underlying dust extraction methodology and data collection are detailed in Mahdavi & Siegel (2020) AS&T (link in "About").