PM2.5 Estimation using Supervised Learning Models
Anusha Anchan1, Manasa G.R.2

1Anusha Anchan, Assistant Professor in the Department of Computer Science & Engineering.
2Manasa G.R., Assistant Professor in the Department of Computer Science & Engineering.
Manuscript received on March 12, 2020. | Revised Manuscript received on March 26, 2020. | Manuscript published on March 30, 2020. | PP: 4771-4776 | Volume-8 Issue-6, March 2020. | Retrieval Number: F9912038620/2020©BEIESP | DOI: 10.35940/ijrte.F9912.038620

Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (

Abstract: Present era of Urbanization, mechanization, and globalization has attracted more and more Air pollution problems. However, PM 2.5 (Particulate Matter) majorly present at air, having diameter below 2.5 μm. With its high concentration leading to health issues such as lung cancer, cardiovascular disease, respiratory disease etc. With respect to this, presented work approach is building of supervised learning models, XGBoost(Extreme Gradient Boosting) along with MLR(Multiple Linear Regression),RF(Random Forest) and MLP (Multilayer Perceptron) to estimate PM2.5 congregation. The air quality data of city Changping in Beijing is taken into consideration for Analaysis. The accuracy of prediction of the four approaches is measured by using contrasting discovered value verses predicted value of PM2.5 with the help of three measuring matrices. The consequences reveals that the Random Forest algorithm outperforms other data mining strategies for the considered data. Prediction of PM2.5 concentrations will assist governing bodies in warning people who are at peak risk, and taking right measures to reduce its quantity in air also to reduce its impact on human life.
Keywords: PM2.5, XG Boost, PM10, Python.
Scope of the Article: Deep learning.