MACHINE LEARNING APPLICATION FOR CARBON ESTIMATION – A CASE STUDY

Climate change is a most global challenging issue. In this regard, a study on carbon dioxide, one of the pollutants causing the climate change is demonstrated in two different states of India viz., Visakhapatnam district (AP) and Shastri Nagar (RJ) for the period of April 2022 to January 2023. Carbon dioxide (CO 2 ) is experiential on hourly and monthly basis for different seasons – summer (April), rainy (July) and winter (December). Most of air pollutants include NO 2 , CO 2 , PM 2.5 , PM 10 etc. that are the major cause for climate change. The air quality in these zones is very poor, highly polluted and risk to humans. The study proved that CO 2 is found comparatively low in rainy season over other. The machine learning regression models were modelled for Visakhapatnam and best models obtained are 1. Step-wise Linear Regression model with MSE (4.51E-28), RMSE (2.12E-14) and R-Squared (1) identified for rainy month. 2. Neural Network Narrow model with MSE (0.462), RMSE (0.680) and R-Squared (0.999) for winter month. 3. Linear Regression model with MSE (0.108), RMSE (0.329) and R-Squared (0.999) for summer. Similarly, the best models for Shastri Nagar for monthly data are 1. Step-wise Linear Regression with RMSE (20.292), MSE (411.774) and MAE (12.524) for April & May (summer), 2. Neural Network Narrow model had RMSE (3.399), MSE (11.554) and MAE (2.141) for July (rainy). 3. Neural Network Bilayered


INTRODUCTION
Global warming is caused by hasty increase in carbon dioxide and other GHG produced by burning oil, gas, coal, fossil, industries and so on.Global warming further causes climate change, one of major issue world-wide.Abnormal climate change brings disaster and harm to humankind (Zhao, 2019).Thereby, global warming is triggering changes in rains, snow patterns and rising sea levels (Letcher, 2021).Thus, one have to build low-carbon and resilient cities to lower the global warming (Lee, 2013& Yoro, 2020) and protect the planet.The huge impact on global warming is due to high emission of CO 2 .The transport and industry sector are significant contributor for CO 2 emission (Lee, 2013& Yoro, 2020) and carbon monoxide (CO) (Wang, 2021).Vehicles produce CO 2 from engine combustion that is harmful carbon monoxide (CO) from incomplete combustion to CO 2 (Wang, 2021).CO 2 and CH 4 are significant emitters for GHG effect (Chapman, 1996).Furthermore, CO poisoning leads to long-term neurological (Rose, 2017); and cardiovascular diseases (Wang, 2021).In atmosphere, high temperature sensitivity results due to weak CO 2 adsorption of pure carbonaceous materials (Kanjilal, 2020).Let's ponder on few remarkable reasons for emission of CO 2.
Firstly, ICT contributes 1.8% to 2.8% of global GHG emissions to environmental pollution, global warming and ultimately climate change.Research indicates the carbon footprint is produced by production and use-phase electricity consumption (Sutton-Parker, 2022).All analysts agree that ICT emissions will not reduce without the determined efforts of political and industrial (Freitag, 2021).
Secondly, carbon credit trading (Chonde, 2016) is for controlling and reducing GHG emission that contributes to global climate change (Robinson, 2006).CO 2 is a concern that happens to be most important GHG produced by combustion of fuels.The other gases emitted from industries include CO 2 , NO 2 , CH 4 and hydro fluorocarbons (Chonde, 2016).Carbon credit allows one ton of CO 2 to be discharged into the air.A great deal of businesses buying/ selling carbon credits for excess emissions has encouraged carbon trading globally.India is among emerging leader for the developing countries in designing innovative strategies and portfolios for carbon trading (Gupta, 2011).From the middle of the 18th century, the concentration of CO 2 with two factors, land use changes and the burning of fossil fuels have been increasing.The higher atmospheric concentrations of CO 2 are absorbed by the oceans (Breeze, 2017).
Thirdly, for amending environmental change, CO 2 capture is considered as the key technology (Kanjilal, 2020).The other is industry penetration of solvent absorption technology for CO 2 captures (Scholes, 2020).The low partial pressure of CO 2 and reduced separation driving force in flue gas applications also hampers the development of membrane technology in industrial applications (Scholes, 2020).
Climate change and air pollution are the biggest environmental concern globally (Coelho, 2021).The study of pollution discusses the negative impacts on environment (Brusseau, 2019).Fourthly, the practice of conservation of forest and soil that enhance the storage of carbon or reduce CO 2 emissions is accomplished with terrestrial carbon sequestration (Sundquist, 2008).

DISCUSSION AND RESULTS
CO 2 remains longer in the atmosphere than any pollutants; this warms the earth causing climate change.Industrial revolution (IR) is related closely to emission of CO 2 causing the pollution.IR increased CO 2 by 50% because of burning fuel, electricity, heat and transportation.Humans have generated an estimated 1.5 trillion tons of CO 2 .CO 2 effects human health including headaches, dizziness, restlessness, breathless, and sudden rise in blood pressure.
In this study of monthly and hourly air quality index (AQI) vs CO 2 and CO for Visakhapatnam (AP) and Shastri Nagar (RJ) for April, July, December and January 2023 were collected from https://app.cpcbccr.com/AQI_India/and ML regression models were developed shown in tables (1,2,3,4).The best models for hourly data are RMSE of SVM-L (5.1594) for April, GPR-SE for July (0.0002) and December (0.0038), and NN-N (0.1932) for January.The response plot of CO 2 vs AQI is shown in figures A.1 in Appendix -A.The model number in each table represents the corresponding graph in Appendix-A and Appendix-B.
For monthly data, the RMSE of best models were LR (0.329) for April & May, SLR model (2.12E-14) for July, NN-N (0.680) for November and December is shown in tables (5 to 7).Visakhapatnam has three seasons with excessive humidity from southern part whereas Shastri Nagar has moderate rains with high heat in summer.
Let's discuss a little on regression models Linear Regression, is a relationship between the response and explanatory parameters, and is represented by where 'y' is the output, constant is denoted by 'b' and 'm 0 ' and slope by 'a' and 'm 1 '.
Stepwise linear regression does multiple regressions a number of times, every time it removes the weakest correlated variable.It is represented by where s xj and s y are the standard deviation for response and explanatory variables.
Treefine/ medium model uses a decision tree to represent how different input variables can be used to predict a target value.Machine learning uses tree-based models for both classification and regression problems.
Support vector machine (SVM) -Linear regression model is a ML tool for classification and regression and is considered a nonparametric technique because it relies on kernel functions.It expressed as Gaussian process regression (GPR) models are utilized for statistical modeling and are nonparametric kernel-based probabilistic models with a finite collection of random variables with a multivariate distribution.
Neural Network model is a computational system that is organized into layers of nodes.

Visakhapatnam models based on the hourly data
It was interesting to note with hourly data on December 2, 2022, the value of CO started with 48 (ppm) at 4:00 am, decrease to 8 (ppm) at 11:00 am, later gradually increased to 36 (ppm) at 9:00 pm.July month had the value of CO to 5 (ppm) on July 24, 2022 with almost < 35 (ppm) the entire month.On April 26, 2022 the highest value of CO (60 ppm) was found with the entire month higher than 31 (ppm).January 6, 2023 had 108 (ppm) at 1:00pm.Hence, it is clear that rainfall washes the pollutant CO in the atmosphere and CO is found less in rainy season.
Visakhapatnam hourly -AQI and CO predictions for April 25 to 30, 2022 Mode l Model Type  The monthly trends of CO and AQI for visakhapatnam are plotted using STATA shown in figure 1 for the months of April, May, July, November and December 2022.AQI and CO seems to be severely bad in December month.
The figure 3 shows the trends of CO (ppm) and AQI for the Visakhapatnam and ShastriNagar from the year 2017 to 2023.AQI seems to be severe in the year January 2019 (AP) whereas AQI is severely bad in the year September 2020 (RJ).
Figure 3.The trends of CO and AQI in AP and RJ.

CONCLUSION
It is very clear that CO 2 is very risky for human health by studying two different statesmoderate and dry climate regions.It is further observed that one need to minimize the emission of CO 2 to protect the planet to attain net zero carbon and let other generations to live on this planet.This regression models will help the researchers, government official and global health organization to make necessary requirements for the highly polluted areas.

Figure 2 .
Figure 2. Trends in Shastri Nagar in ShastriNagar from the year 2017 to 2023

Table 5 .
Regression models for April and May monthVisakhapatnam monthly-AQI and CO predictions for July 2022

Table 8 .
Regression models for April month The global roadmap is set for 2025 to have no new coal plants and by 2050 to have clean green energy sources and affordable energy, to achieve the net carbon zero globally.