Research Project (Jan, 2021 - April 2021)

• COURSE CODE: CSE498R (Directed Research)
  
• Research Supervisor: Dr. Shazzad Hosain
  Professor
  Department of Electrical and Computer Engineering
  North South University (NSU)

Authors

• Rutton Chandra Sarker

• Aniket Banik

Prediction of Air Quality Index(AQI) in Bangladesh Using Machine Learning Techniques

• In today’s modern world, air pollution is an intense problem because it is the only reason behind the harmful effects not only on human health but also on all living conditions. Air pollution-related deaths and diseases have been on the increase in recent years, according to available surveys and research, exacerbating people's suffering, especially children, women, and the elders. This reserach based project addresses the prediction of air quality index (AQI) level based on various atmospheric condition observed in Dhaka, Khulna and Rajshahi City of Bangladesh. We build a ML model that predicts AQI level in Bangladesh using supervised machine learning algorithms. This ML model will bring benefit for government to take legal actions for reducing the air pollution at moderate level.

Library Used

- Pandas
 
- NumPy
 
- matplotlib
 
- Seaborn

Data Collection

  - Weather Dataset (Independent Features)

      * Dhaka(2016-2020)
  
      * Khulna (2018-2020)
  
      * Rajshahi (2018-2020)

 
  - Air quality Index Dataset (Dependent Features)
  
 ------------------------------------------------------------
 |  AQI Range  |       Category            |    Color       |
 |----------------------------------------------------------|
 |   0-50      |   Good(G)                 |  Green         | 
 |   51-100    |   Moderate(M)             |  Yellow Green  | 
 |   101-150   |   Caution(C)              |  Yellow        | 
 |   151-200   |   Unhealthy(U)            |  Orange        |
 |   201-300   |   Very Unhelthy(VU)       |  Red           |
 |   301-500   |   Extremely Unhealthy(EU) |  Purple        |
 ------------------------------------------------------------

ML Algorithms Applied

      - Linear Regression
      
      - Lasso Regression
      
      - Ridge Regression
      
      - KNN Regressor
      
      - Decision Tree Regressor
      
      - Random Forest Regressor
      
      - Xgboost Regressor