Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Team9_ITBgg

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Predictive Data Analysis

of
CO2 emissions

Data Mining and Predictive Analysis Lab (ICT 3262)


Dept. of Information & Communication Tech.

Tarun Pakhare :200911208


Rokkam Sai Kiran Reddy:200911230

Date of submission : 04/04/2023


INTRODUCTION:

CO2 emissions have become a massive problem for environmentalists, policymakers, political leaders, and
the public. CO2 emissions are primary contributors to Global Warming and Climate change. We realize that
increasing emissions will only get more hazardous for us. With the ever-growing talk around climate change
and its effects, it's necessary to understand the need to reduce emissions or start planning the way forward.
As a result, there is a need for growing research and analysis of CO2 emissions so we can predict future
emissions and plan how to reduce them.

In this project, we aim to analyze two datasets related to CO2 emissions based on the country using four
different algorithms. We plan on analyzing the data using these algorithms and calculating each algorithm's
precision. We will be comparing the various algorithms and exploring the relationship between CO2
emissions and the following factors:

• Cereal Yield
• Protected Area Percentage
• Energy used per capita.
• Percentage of the population living in the urban sector
• Percentage of population growth

Ultimately, we plan to provide the best model to predict CO2 emissions and provide an effective a plan of
action to lessen the effects of climate change.

METHODOLOGIES:
Random Forest Algorithm:

It is a popular machine-learning algorithm which combines the output of various decision trees to produce a
single result. We can break down the Random Forest Algorithm into the following steps:

Step-1: Defining the problem: We start by defining the problem as “Prediction of CO2 emissions based on
few factors”.

Step-2: We use the following dataset: “https://datacatalog.worldbank.org/dataset/climate-change-data”


and clean the data ensuring that there are no outliers in the data.

Step-3: Split the data: We split the data into training set and testing set in the ratio 3:7. The training set
helps in building the model while the testing set helps in evaluating the performance of the model.

Step-4: Choose Hyperparameters: We set the hyperparameters such as:


• n_estimator_desc_trees: Number of decision trees
• max_features_tree_split: Number of features to consider at every split.
• max_depth_leaves_tree: Maximum number of leaves in a tree
• min_sample_split_node: minimum samples required to split a node.
• Min_sample_at_leaf: minimum number of samples required at each leaf.
Step-5: Train the model: Fit the training data to RFA. We then train the model using a random subset and
with cross-validation. We obtain a R2 score.

Step-6: Evaluate the model: In this step we evaluate the model we created using the testing set. We obtain
another R2 score which we compare with the R2 score obtained before.

Step-7: Validate the model: We use cross-validation techniques to check if the model is still able to
generalize new data.

Results:
R2 obtained using training set = 0.959
R2 obtained using testing set = 0.958
We observe that that the test set shows a promising correlation with the training set.

Linear Regression:
It is a Machine Learning Algorithm used for supervised learning. It predicts a dependent variable based on
the independent variables. We can break down the Linear Regression into following steps:

Step-1: Defining the problem: We start by defining the problem as “Prediction of CO2 emissions based on
few factors”.
Step-2: We use the following dataset: “https://datacatalog.worldbank.org/dataset/climate-change-data”
and clean the data ensuring that there are no outliers in the data.

Step-3: Choosing a Linear Regression model: We are using the Simple Linear Regression Model.

Step-4: Fit the data: We fit the data to the model. This involves finding the coefficients of the line that best
fits the data.

Step-5: In this step we create the training and testing datasets in the ratio 3:7.

Step-6: Evaluating the model: We evaluate the model using the training and testing datasets and calculate
the coefficient of determination(R-squared).

Step-7: Validate the model: We use cross-validation techniques to check if the model is still able to
generalize new data.

Results:
We get a value of 0.91 for R2 using this algorithm
Literature review

CO2 Emissions from Renewable and Non-Renewable Electricity Generation


Sources in the G7 Countries: Static and Dynamic Panel Assessment: (Published
2023)

The article provides the issue of global warming and its detrimental effects. The writer discusses various
factors such as industrialization, urbanization, population expansion, and lifestyle changes.
The writer says a significant contributor to the problem is the emission of carbon dioxide.
It also provides analysis by applying Various statistical techniques. The analysis gives a positive correlation
between CO2 emissions and energy production from coal and gas.CO2 emissions can be reduced using Coal-
fired power plants, Hydroelectric and renewable energy production, and nuclear energy. The study
concludes that reducing CO2 emissions is possible by moving towards renewable and hydroelectric sources,
and it suggests several policies for sustainable energy.

A review of trends and drivers of greenhouse gas emissions by sector from 1990 to 2018
(Published 29 June 2021)
This article discusses the sources and trends of global greenhouse gas (GHG) emissions in five key economic
sectors including and not limited to energy and transport. The review estimates GHG emissions trends,
highlighting the significant sources of emissions. In Europe and North America, moderate decarbonization
of energy systems has been achieved through fuel switching and renewable energy adoption. This review
underscores the importance of identifying, understanding, and addressing the continuous climate change
trends across sectors as humanity navigates the Anthropocene era.

Investigating the impact of multi-dimensional urbanization and FDI on carbon emissions in


the belt and road initiative region: (Published 2023)
This article discusses the international organizations becoming increasingly concerned about how
urbanization may affect carbon emissions, this study says the direct and indirect effects of ever-growing
urbanization on carbon emissions in the BRI region were quantified using multiple statistical approaches
and remote sensing data sets. Results says that East and South Asian countries reporting the highest
increases in carbon emissions as well as the fastest rates of multi-dimensional growth and foreign direct
investment. The results also highlight the significance of including emission-reduction measures in
population, economic, land use, and foreign investment policies as doing so will strengthen the
collaboration among global BRI nations in managing carbon emissions.

Climate and socioeconomic drivers of biomass burning and carbon emissions from fires in
tropical dry forests: (Published 2022)

The article discusses the global decline in the burned area over the years, with drylands accounting for a
significant proportion of the fires. While savannas have seen a reduction in the burned area, tropical dry
forests (TDF) have experienced an increase in fire incidence, which negatively impacts environment and
increases carbon emissions. Climate change will only make matters worse, with TDF predicted to expand,
increasing the risk of fires. Yet there have yet to be many studies to investigate the impact of climate and
socioeconomic factors on TDF fire regimes. The findings highlight the need for improved fire management
to reduce the negative impact on biodiversity and carbon emissions and achieve the goal of "half protected
species" globally.

Nuclear energy consumption and CO2 emissions in India: (Published 2022)

This study analyzes the relationship between nuclear energy use and CO2 emissions in India, using data
from 1970 to 2016 and the Fourier Autoregressive Distributive Lag model. The study notes that India's
rapidly growing population is placing a significant strain on the environment. However, with 22 operational
nuclear reactors, India has great potential to reduce CO2 emissions. The study validates India's
Environmental Kuznets Curve (EKC) hypothesis, finding that the significant coefficients of GDP and GDP.2
support it. The study also suggests that most environmental externalities are corrected within a year.
Considering these findings, the study suggests policymakers support the expansion of nuclear power, enact
new ecological laws that support responsible atomic energy use, and reduce dependence on fossil fuels.
Finally, the study suggests that future research should explore nonlinearities in the nuclear energy-CO2
emissions relationship.

A new framework to the green economy: (Published 2022)

The study examined the connection between energy sector investments made through public-private
partnerships and CO2 emissions in several Asian economies between 1990 and 2019. The link is estimated
by the study using the NARDL method. The findings show that different countries experience different
effects of PPPIE on CO2 emissions. Long-term CO2 emissions in Indonesia and Russia rise as a result of a
favorable change in PPPIE. Moreover, it has a detrimental impact on CO2 emissions in Turkey and China.
The research suggests that governments subsidize the renewable energy industry and create public-private
partnerships in order to enhance environmental quality.

Green Technology Innovation, Globalization, and CO2 Emissions: (Published 2021)

This investigation looks at the connection between various innovations, globalization, and CO2 emissions in
countries. Advanced econometric methods are used, and an interaction variable (TIGLOB) is included to
assess the moderation effects of the factors mentioned above. The findings reveal a negative and significant
relationship between technological innovation and CO2 emissions in all regions. Globalization has a positive
impact on CO2 emissions in OBOR and South Asia but a negative impact in few other regions. The
interaction effect of TIGLOB is significantly negative for OBOR, but also positive for few countries. The study
is robust, and the findings have implications for policy and recommendations for reducing CO2 emissions.

Greenhouse gases utilization: A review (Published 2021)

This article/paper comprehensively reviews the excessive global emissions of Co2 and greenhouse gases. It
emphasizes the effects of greenhouse gases and how they must be contained. The authors begin by
explaining the ill effects of the emissions and how fossil fuels take millions of years to form. They discuss the
ways in which emissions could be reduced and feel that a “Carbon Capture and Utilization” (CCU) could be
the best solution as it would ensure the capturing as well as re-utilizing of CO2 and CH4 gases. They also
mention the four different processes by which greenhouse gases and be converted into valuable products.
The authors conclude by speaking about the barriers to producing CCU’s which include factors such as
economic conditions and poor electrical conductivity, among others.
Overall, this article gives an insight into greenhouse gases and their effects on human beings and provides
solutions for their containment of them. It also speaks about the barriers and issues which need to be
resolved in the future.

The influence of political ideology on greenhouse gas emissions. (Published 2022)

This article/paper provides a systematic review of the influence that political ideologies have on greenhouse
gases. This paper covers a dataset from the year 1990 to 2016 covering 98 different countries taking CO2,
CH4, and N2O emissions as the basis for greenhouse gas emissions. The authors analyze the data and
include various decisions taken by the left-wing parties as well as right-wing parties. The authors provide
their analysis by stating that “Leftist parties increase energy efficiency and spend more on secondary
education, which leads to less greenhouse gas emission.” They conclude that governments should initiate
more measures to reduce greenhouse gas emissions. Some of these include improving the educational
system as well as spreading awareness to the people about the effects of CO2 emissions.

Countries with sustained greenhouse gas emissions reductions: analysis of trends and
progress by sector (Published 2021)

This article/paper provides a review of the countries which have managed to reduce/sustain greenhouse
gases over a few years. The authors begin by talking about greenhouse gases and their effects over the
years. They explore various sectors, which include but are not limited to electricity production and
transportation. They analyze the multiple changes which have taken place and the factors that have caused
this. They also analyzed the economic growth along with the reduction of emissions, bridging the
similarities between both. They conclude by stating that many more countries are going to fast achieve the
reduction of emissions, and more substantive policies would help in reducing the emissions. They also
report that the most reduction of emissions is in the energy sector while the transportation sector
continues to rise.

Forecasting of future greenhouse gas emission trajectory for India using energy and
economic indexes with various metaheuristic algorithms(Published 2022)

This article/paper provides a systematic review of greenhouse gas emissions and their forecast over a
period of a few years. The authors begin by addressing the greenhouse effect levels and their impact. India
is one of the highest producers of greenhouse gases, so it is essential to identify its trend to add policies
that help in reducing emissions. The authors use various algorithms to identify this trend as it has a global
effect. The authors analyze the data from 1990 to 2018 and, on its basis, produce various expectancies of
CO2, N2O, and F-gases for the year 2050. The authors identify that Marine Predators Algorithm (MPA) is the
best way to determine the forecast. They conclude by stating that new policies should be implemented to
reduce energy consumption or look for different alternatives to reduce the currently increasing Greenhouse
gas emissions.
A Hybrid Model with Applying Machine Learning Algorithms and Optimization Model to
Forecast Greenhouse Gas Emissions with Energy Market Data (Published 2022)

This article/paper provides a review of building a model for forecasting Greenhouse gas emissions in Iran.
The authors begin by collecting Greenhouse gas emission data from 1990-2018 in the Energy sector. They
have emphasized the energy sector as it continues to be the top-most producer of greenhouse gas
emissions across the world. The authors have evaluated the data using nine different algorithms whose
accuracy has been evaluated with five different indicators. They have predicted the emissions up to the year
2028 in this paper. The authors conclude by stating that the hybrid model could predict more accurately
than most machine learning algorithms. They found a result that the emissions in Iran could be 1096
Mt/year by 2028.

Future greenhouse gas emissions from metal production: gaps and opportunities towards
climate goals (Published 2022)

This article/paper provides a systematic review of the emissions from production of metal and how they can
be managed. The authors begin by estimating the production of metal for 2010-2100. Once they are able to
get the estimates of metal that would be required, they then proceed towards the effect of greenhouse gas
emissions due to the production of these metals. The authors have focused on six metals and the factors
that affect their production as well as their emissions. The authors have done a thorough decomposition
and sensitivity analysis to reduce future greenhouse gas emissions. They conclude by stating that going by
the current trends, emissions from metal production would not be able to maintain the set sustainability
levels. Hence, it is important to have goals set to reduce greenhouse gas emissions as well as to make the
existing metals sustainable to reduce the production of metals as well as improve the recycling rate.

Mitigating greenhouse gas emissions from irrigated rice cultivation through improved
fertilizer and water management (Published 2022)

This article/paper analyses the effects of greenhouse gases released by Agricultural activities and provides
solutions to them. The authors begin by specifying the greenhouse effect due to agricultural activities. They
mention a few solutions which could be followed to mitigate the greenhouse effect. The authors mention
conducting experiments in the fields that gave promising results. The authors conclude by stating that even
though greenhouse emissions cannot be reduced to zero, there are ways to reduce them by increasing
Nitrogen usage.

Evaluation of greenhouse gas emissions from the European urban wastewater sector, and
options for their reduction. (Published 2022)

This article/paper evaluates the effect of greenhouse gas emissions in Europe due to wastewater. The
authors present the number of emissions done due to wastewater and give a few solutions for the same.
They cover six different factors causing emissions. They get an estimate of the emissions released. They
conclude that water treatment plants could reduce emissions by making the right use of electricity and its
decarbonization.
REFERENCES:
CO2 Emissions from Renewable and Non-Renewable Electricity Generation Sources in the G7 Countries:
Static and Dynamic Panel Assessment by Liton Chandra Voumik 1ORCID,Md. Azharul Islam 1,Samrat Ray
2,Nora Yusma Mohamed Yusop 3,*ORCID andAbdul Rahim Ridzuan 4,5,6,7,8,*ORCID

A review of trends and drivers of greenhouse gas emissions by sector from 1990 to 2018
William F Lamb30,1,2, Thomas Wiedmann3, Julia Pongratz4,5, Robbie Andrew6, Monica Crippa7, Jos G J
Olivier8, Dominik Wiedenhofer9, Giulio Mattioli2,10, Alaa Al Khourdajie11, Jo House12Show full author list

Investigating the impact of multi-dimensional urbanization and FDI on carbon emissions in the belt and road
initiative region: Direct and spillover effects Author links open overlay panelGuoen Wei a b, Mo Bi b, Xiao
Liu c, Zhenke Zhang b, Bao-Jie He d e f g

Climate and socioeconomic drivers of biomass burning and carbon emissions from fires in tropical dry
forests: A Pantropical analysisRogelio O. Corona-Núñez1,2,3| Julio E. Campo

Nuclear energy consumption and CO2 emissions in India: Evidence from Fourier ARDL bounds test approach
Author links open overlay panelOnder Ozgur a, Veli Yilanci b, Maxwell Kongkuah c

A new framework to the green economy: asymmetric role of public-private partnership investment on
environment in selected Asian economies
Bo Gao,Ilhan Ozturk &Sana Ullah

Green Technology Innovation, Globalization, and CO2 Emissions: Recent Insights from the OBOR Economies
by Ahmer Bilal 1,*,Xiaoping Li 1,Nanli Zhu 2,*,Ridhima Sharma 3 andAtif Jahanger 4

Greenhouse gases utilization: A review


Luqman Jeffry a f, Mei Yin Ong a, Saifuddin Nomanbhay a, M. Mofijur b c, Muhammad Mubashir d, Pau Loke
Show e

The influence of political ideology on greenhouse gas emissions


Quan-Jing Wang a, Gen-Fu Feng b, Hai-Jie Wang a, Chun-Ping Chang c

Countries with sustained greenhouse gas emissions reductions: ananalysis of trends and progress by
sectorWilliam F. Lamba,b, Michael Grubbc, Francesca Diluisoaand Jan C. Minx

Forecasting of future greenhouse gas emission trajectory for India using energy and economic indexes with
various metaheuristic algorithms Hüseyin Bakır a, Ümit Ağbulut b, Ali Etem Gürel b c d, Gökhan Yıldız b,
Uğur Güvenç e, Manzoore Elahi M. Soudagar f g, Anh Tuan Hoang h, Balakrishnan Deepanraj i j, Gaurav Saini
g, Asif Afzal f k l m

Machine Learning Algorithms and Optimization Model to Forecast Greenhouse Gas Emissions with Energy
Market Data Emami Javanmard a b, S.F. Ghaderi a b

Future greenhouse gas emissions from metal


production: gaps and opportunities towards
climate goals†
Ryosuke Yokoi, *a Takuma Watari bc and Masaharu Motoshita a

Mitigating greenhouse gas emissions from irrigated rice cultivation through improved fertilizer and water
management
Author links open overlay panelS.M. Mofijul Islam a, Yam Kanta Gaihre b, Md. Rafiqul Islam a, Md. Nayeem
Ahmed a, Mahmuda Akter c, Upendra Singh b, Bjoern Ole Sander d

Evaluation of greenhouse gas emissions from the European urban wastewater sector, and options for their
reduction
Author links open overlay panelVanessa Parravicini a, Per Henrik Nielsen b, Dines Thornberg c, Alberto
Pistocchi d

You might also like