Faculty of Computing and Information Technology (FCIT)

Department of Computing Indus University, Karachi

NAME OF STUDENT: Danial Riaz ID No: 1951-2021

Assignment-3
SPRING 2025
Department Faculty Total
Program Semester Course Title Instructor Issue Date Due Date
/ Faculty Signature Marks
BS 8
Machine Hira Shafi
FCIT Software
Learning
11/06/2025 18/06/2025 5
Engineering

Instructions
1. This assignment contains 1 Question. Attempt all questions.
2. This assignment should be completed within the in assigned time, after the due time, assignment
will not be accepted.
3. Students of particular course will download assignment and submit solution which will only be
accepted through CMS portal.
4. Please note that copy-paste is a serious nature of academic dishonesty, it is called "Plagiarism"
and the penalties are attached to being found guilty of committing such offenses.
5. It is allowed using lecture notes, books, and other sources, however needing to refer/cite properly.
Reference list must be given at end of the assignment.
6. This assignment should be submitted in PDF file for this purpose first take image of all hand
written pages and then merge using Smartphone app (from PC/Laptop put all images in word file
and save as PDF) including assignment paper in the start of submission.
7. Assignment can be compressed or break in two parts if file size is larger than uploading limit.
8. The font size should be 12 in Times New Roman, Line Space 1.5, Main Heading font size Bold
14, Subheading font size 12 Bold and text font size 12 should be used. All figures and
illustrations should be properly titled or numbered on the left side, below.
9. Also ensure that no part of your assignment has been written by any other person, except to the
extent of collaboration and /or group work.
10. In case of theory, the assignment should not exceed than 1000 words and not less than 600 words.
11. The expected outcome of this assignment follows (C3, PLO-5, CLO-3)
This table is for official use; do not write anything on it.

CLOs CLO_3 Total

Question 1 1
Number

Student’s Score

Maximum Score
5 5

This paper has a total of 2 pages including this title page

 Faculty of Computing and Information Technology (FCIT)
Department of Computing Indus University, Karachi

NAME OF STUDENT: Danial Riaz ID No: 1951-2021

Assignment # 3

SPRING 2025
Machine Learning
Expected Outcome: PLO-5 (C3 - Apply)
CLO-3: Apply various techniques and algorithm to model the real world problems into machine
learning techniques

Total (5 marks)
5 marks
Question NO 01: Based on the scenarios given, share your justification as possible:

A: Predicting Student Exam Outcomes

A high school wants to develop a model that can predict whether a student is likely to pass or fail the final
exam. The school has access to data from previous years, including:
• Hours studied per week
• Attendance rate (%)
• Previous exam scores
• Participation in extra tutorials (Yes/No)
• Final exam result (Pass/Fail)
The goal is to train a model using this historical data so that it can predict future outcomes for new students.

Answer the following:

1. What is the objective of this task? Is it classification, prediction, or pattern discovery?

2. What type of data is available? Are the outcomes labeled or not?
3. Based on the data and the objective, which type of machine learning model would you use and
why?
4. How would you evaluate the model’s performance? (e.g., accuracy, confusion matrix)
5. Can you name at least one algorithm that fits this type of problem? Why is it suitable?
 Faculty of Computing and Information Technology (FCIT)
Department of Computing Indus University, Karachi

NAME OF STUDENT: Danial Riaz ID No: 1951-2021

B: Identifying Learning Behavior Patterns

An online learning platform wants to better understand its users and customize their learning experience. The
platform collects the following behavioral data:
• Average time spent per session
• Number of sessions per week
• Quiz performance scores (normalized)
• Number of interactions on the discussion forum
• Resources downloaded
There is no existing label or category for each user. The goal is to group students into clusters based on
similar learning patterns to create personalized learning paths.

Answer the following:

1. What is the objective of this task? Are we predicting something or discovering patterns?
2. Are there any labels or target outcomes in the dataset?
3. Based on the task and the nature of the data, which type of machine learning approach would
you choose and why?
4. How would you evaluate the quality of the results? (e.g., silhouette score, visual cluster analysis)
5. Can you name at least one algorithm that fits this kind of problem? Why is it appropriate?
 Faculty of Computing and Information Technology (FCIT)
Department of Computing Indus University, Karachi

NAME OF STUDENT: Danial Riaz ID No: 1951-2021

A: Predicting Student Exam Outcomes

1. What is the objective of this task?

Objective: The goal is to predict a categorical outcome (Pass or Fail), so this is a classification task.

2. What type of data is available?

The data is labeled we have input features (e.g., hours studied, attendance rate) and a known outcome (Pass/Fail).
This means it's asupervised learning problem.

3. Which type of machine learning model would you use and why?

A supervised classification model is appropriate because:

The outcome is binary (Pass/Fail).

Historical data with known outcomes is available for training.
The goal is to make predictions on unseen students based on similar features.

4. How would you evaluate the model’s performance?

You can evaluate the model using:

Accuracy: Percentage of correct predictions.

Confusion Matrix: To understand the count of true positives, true negatives, false positives, and false negatives.
Precision & Recall: Useful if class imbalance exists (e.g., far more passes than fails).
F1 Score: Harmonic mean of precision and recall.

5. Can you name at least one algorithm that fits this type of problem?

Logistic Regression is a good fit because:

It handles binary classification problems well.

It is simple and interpretable.
It works effectively with numerical and categorical features (e.g., hours studied and participation in tutorials).

Other possible algorithms: Decision Tree, Random Forest, or Support Vector Machine (SVM).

B: Identifying Learning Behavior Patterns

1. What is the objective of this task?

Objective: The task aims to discover hidden patterns or groups in user behavior, so this is a case of unsupervised
learning, specifically clustering.

2. Are there any labels or target outcomes in the dataset?

No, the dataset does not include labels or predefined categories for each user. It only contains behavioral metrics.

3. Which type of machine learning approach would you choose and why?

An unsupervised clustering approach is appropriate because:

We want to discover natural groupings in the data.

There are no labels to guide the model during training.
Clustering helps segment users based on behavioral similarities.
 Faculty of Computing and Information Technology (FCIT)
Department of Computing Indus University, Karachi

NAME OF STUDENT: Danial Riaz ID No: 1951-2021

4. How would you evaluate the quality of the results?

Evaluation methods include:

Silhouette Score: Measures how similar an object is to its own cluster compared to others.
Elbow Method: Helps determine the optimal number of clusters (k).
Visual Cluster Analysis: Using PCA or t-SNE to reduce dimensions and visually assess cluster separation.

5. Can you name at least one algorithm that fits this kind of problem?

K-Means Clustering is suitable because:

It is effective for grouping data into k clusters based on feature similarity.

It works well with numerical data like time spent, quiz scores, etc.
It’s relatively easy to implement and interpret.

Other suitable algorithms: DBSCAN, Hierarchical Clustering, Gaussian Mixture Models (GMM).