Scribd
Upload
10 views4 pages

Assignment 1 (Fall 2024)

Uploaded by

20208046
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
10 views4 pages

Assignment 1 (Fall 2024)

Uploaded by

20208046
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Cairo University

Faculty of Computers and Artificial Intelligence


Machine Learning Course

Assignment 1: Linear and Logistic Regression


Most vehicles on the road today use petrol or diesel fuel engines. When these
fuels burn, they release energy that powers the vehicle. However, this
process also produces waste in the form of carbon dioxide gas (CO2). The
amount of CO2 emission is a basic indication of a vehicle's impact on the
environment and air quality. So, with climate change at the forefront of global
issues, it is crucial to understand and predict the vehicles’ CO2 emission
amounts.

In this assignment, you are required to build a linear regression model and a
logistic regression model to predict the amounts of CO2 emission and their
respective classes based on the vehicles’ data.

Dataset:
The attached dataset “co2_emissions_data.csv” contains over 7000 records
of vehicles' data with 11 feature columns in addition to 2 target columns. The
features are: the vehicle’s make, model, size, engine size, number of cylinders,
transmission type, fuel type, and some fuel consumption ratings columns.
The targets are the CO2 emission amount (in g/km) and the emission class.
Note: This dataset is a modified version of the "CO2 Emission by Vehicles”
dataset. The original dataset was obtained from Kaggle.
Requirements:
Write a Python program in which you do the following:
a) Load the "co2_emissions_data.csv" dataset.
b) Perform analysis on the dataset to:
i) check whether there are missing values
ii) check whether numeric features have the same scale
iii) visualize a pairplot in which diagonal subplots are histograms
iv) visualize a correlation heatmap between numeric columns
c) Preprocess the data such that:
i) the features and targets are separated
ii) categorical features and targets are encoded
iii) the data is shuffled and split into training and testing sets
iv) numeric features are scaled
d) Implement linear regression using gradient descent from scratch to
predict the CO2 emission amount.
-> Based on the correlation heatmap, select two features to be the
independent variables of your model. Those two features should have a
strong relationship with the target but not a strong relationship with
each other (i.e. they should not be redundant).
-> Calculate the cost in every iteration and illustrate (with a plot) how
the error of the hypothesis function improves with every iteration of
gradient descent.
-> Evaluate the model on the test set using Scikit-learn’s R2 score.
e) Fit a logistic regression model to the data to predict the emission class.
-> Use the two features that you previously used to predict the CO2
emission amount.
-> The logistic regression model should be a stochastic gradient
descent classifier.
-> Calculate the accuracy of the model using the test set.
Remarks:
● You can use functions from data analysis and computing libraries (e.g.
Pandas and NumPy) as you please throughout the entire code.
●
● You can use machine learning libraries such as Scikit-learn for
preprocessing and metrics but NOT for "from scratch" requirements.

● The train/test split has to be performed before the feature scaling step.
●
● The numeric features of the test set should be scaled using the statistics
of the train set that were used to scale it.
●
● You should use the R2 score to evaluate the linear regression model as it
provides a measure of how well observed outcomes are replicated by
the model. In general, the best possible score is 1, but the score can be
negative as the model can be arbitrarily worse.

Deliverables:
● You are required to submit ONE zip file containing the following:
○ Your code (.py) file.
If you have a (.ipynb) file, you have to save/download it as (.py)
before submitting.
○ A report (.pdf) containing the team members' names and IDs, and
the code of each requirement with screenshots of the output of
each part.

● The zip file MUST follow this naming convention:


Group_A1_ID1_ID2_ID3_ID4
Submission Remarks:
● The maximum number of students in a team is 5 and the minimum is 3.
● Team members must be from the same lab (or have the same TA).
● All team members must understand all parts of the code.
● No late submission is allowed.
● Stick to uploading ONLY the required files following the naming
convention: Group_A1_ID1_ID2_ID3_ID4
● A penalty will be imposed for violating any of the assignment rules.
● Cheaters will get ZERO and no excuses will be accepted as per the
“Plagiarism Scope” document.

Grading Criteria:
Both the code and the report must include:
Analysis 4 marks (1 mark per step)
Preprocessing 4 marks (1 mark per step)
Linear Regression
Selected features 1 mark
Gradient descent 3 marks
Cost function and plot 1 mark
2
R score 1 mark
Logistic Regression
Classifier 3 marks
Accuracy 1 mark
The total is 18 marks (will be scaled to 6 marks)

You might also like