Pokhara University

Faculty of Science and Technology

Course Code.: CMP 336 Full marks: 100

Course title: Data Science and Machine Learning (3-1-2) Pass marks: 45
Nature of the course: Theory & Practical Time per period: 1 hour
Year, Semester:…………… Total periods: 45
Level: Bachelor Program: BE Software

1. Course Description
This course provides a comprehensive introduction to the fields of Data Science and Machine
Learning, aimed at equipping students with the essential knowledge and practical skills required
to analyze data, interpret data, apply machine learning methods and visualize results.

It will include the following information:

● Covers a wide range of topics, including data pre-processing, statistical analysis, machine
learning algorithms, model evaluation, and the application of these techniques to solve real-
world problems.
● Delivery approach includes hands-on labs, case studies and deep understanding of how to
leverage data to make informed decisions.

2. General Objectives
The course is designed with the following general objectives:
• To provide students with a foundational understanding of Data Science and Machine
Learning.
• To familiarize students with techniques for cleaning, transforming, and visualizing data to
uncover patterns and insights.
• To provide the knowledge for use of mathematics such as statistics, probability for data
analysis and machine learning,supervised learning algorithms, linear regression, decision trees,
and support vector machines, and their applications.

• To expose students to unsupervised learning techniques such as clustering and

dimensionality reduction, and their use in identifying patterns and simplifying data.
.
3. Contents in Detail
This section contains the details to be taught under the course.
Specific Objectives Contents

● Intends to provide a brief Unit I: Introduction to Data Science and Machine

introduction to the field of Data Learning (4Hrs)
Science and Machine Learning. 1.1 Definition and Overview of Data Science and
● Learn about various domains Machine Learning
within Data Science and how they 1.2 Applications of Data Science in various industries
interrelate. 1.3 Types of Data: Clean Data and Dirty Data
● Helps students understand the 1.3 Data Science, AI, and Machine Learning
significance of data in modern
decision-making.
● Intends to get students well- Unit II: Data Collection and Preprocessing (7 Hrs)
acquainted with data collection 2.1 Different Data Collection Methods for Machine
methods and preprocessing techniques. Learning: Surveys, Sensors, Web Scraping, APIs,
● Able to apply various Databases
preprocessing techniques to clean and 2.2 Data Quality Issues: Missing Data, Noisy Data,
prepare data for analysis. Inconsistent Data, Data Transformation
2.3 Techniques for Handling Missing Data
2.4 Data Cleaning Techniques: Handling Outliers,
Dealing with Categorical Data, Normalization, and
Standardization
2.5 Dependent and independent variables
● Intends to provide students with Unit III: Exploratory Data Analysis (6Hrs)
the skills to explore and understand 3.1 Introduction to EDA
data. 3.2 Descriptive Statistics: Mean, Median, Mode, Standard
● Learn about various EDA Deviation, Variance, Skewness, Kurtosis
techniques to identify patterns and 3.3 Data Visualization Techniques: Histograms, Box
insights in the data Plots, Scatter Plots, Heatmaps
3.4 Identifying Trends: Mann–Kendall, Spearman's Rank,
Sen’s Slope
3.5 Correlations
3.6 Introduction to Hypothesis Testing
● Intends to provide students with Unit IV: Data Engineering (5 Hrs)
the skills to explore and understand 4.1 Data pipeline, Design and Monitoring
data. 4.2 Extract, Transform and Load (ETL)
● Learn about various EDA 4.3 Feature Engineering
techniques to identify patterns and 4. 5 Feature Selection
insights in the data 4.6 Dimensionality Reduction: PCA, LDA

● Helps students learn how to Unit V: Introduction to Machine Learning (9 Hrs)

implement basic machine learning 5.1 Definition and Types of Machine Learning:
models. Supervised, Unsupervised Learning, Reinforcement
● Able to differentiate between Learning
various machine learning algorithms 5.2 Overview of the Machine Learning
and their applications.
 5.3 Supervised Learning: Linear Regression, Logistic
Regression, Decision Trees, Random Forest, k-NN,
Support Vector Machines (SVM)
5.4 Unsupervised Learning: k-Means Clustering,
Hierarchical Clustering methods
Key Concepts: Training, Testing, Validation, Overfitting,
Underfitting
● Apply basic and machine Unit VI: Anomaly Detection (4 Hrs)
learning methods for detecting 6.1 Definition
anomalies. 6.2 Types: point, contextual, collective
6.3 Applications
6.4 Techniques for Anomaly Detection
6.4.1 Statistical Methods
6.4.2 Distance-based Methods
6.4.3 Density-based Methods
6.4.4 Clustering based Methods
6.4.5 Common Methods (one-class classification,
isolation forest)
6.5 Anomaly Detection in High-Dimension
● Intends to get students well- Unit VII: Model Evaluation and Optimization (6 Hrs)
acquainted with model evaluation 7.1 Confusion Matrix,
techniques. 7.2 Evaluation Metrics
● Able to make use of various 7.2.1 Supervised: Accuracy, Precision, Recall, F1
optimization techniques to improve Score, ROC Curve, AUC, MSE, True Positive Rate, False
model performance. Positive, MSE, MAE, RMSE
7.2.2 Unsupervised: Purity, Rand Index, Silhouette
Coefficient, Dunn Index
7.3 Cross-Validation Techniques
7.4 Hyperparameter Tuning: Grid Search, Random Search
7.5 Model Selection Techniques: Bias-Variance Trade-
off, Ensemble Methods (Bagging and Boosting)
7.6 SMOTE Technique to Handle Imbalance
7.7 Time & Space Complexity of Machine Learning
Models
● Helps students understand the Unit VIII: Ethical and Legal Considerations in Data
ethical implications and legal Science (4 Hrs)
considerations in data science. 8.1 Data Privacy and Security
8.2 Ethical Issues in Data Science: Bias, Transparency,
Accountability
8.3 Legal Considerations: Data Protection Laws,
Intellectual Property

4. Methods of Instruction
The course will utilize a mix of lectures, tutorials, case studies, and lab sessions to support
learning. Lectures will deliver core knowledge, while tutorials and case studies will enhance
comprehension. Lab sessions will provide hands-on experience, enabling students to apply
theory to practical, real-world situations. This integrated approach ensures a well-rounded
learning experience, fostering both theoretical insight and practical skills essential for success in
data science and analytics.

5. Case Studies
Students will complete the following case studies and submit their reports:
● Exploratory Data Analysis (Agricultural Commodities): Students will conduct a
comprehensive exploratory data analysis on a dataset related to agricultural commodities. This
will involve analyzing trends, patterns, and correlations to provide insights.
● Supervised Learning (Customer Churn Prediction in Telecommunications): Students will
build and evaluate a supervised learning model to predict customer churn in the
telecommunications industry. The case study will require them to preprocess data, select relevant
features, and apply classification algorithms to identify customers at risk of leaving.
● Anomaly Detection in Real-World Applications: Students will implement anomaly
detection techniques to identify unusual patterns or outliers in a real-world dataset. This case
study will involve applying various anomaly detection methods to solve practical problems such
as fraud detection or system monitoring.
Students are required to submit a detailed report documenting their approach, results, and
analysis.

6. List of Tutorials

The following tutorial activities of 15 hours per group of maximum 24 students should be
conducted to cover all the required contents of this course.

S.N. Tutorials
1 ● Using libraries of your programming choices (e.g. pandas, R) to
manipulate datasets.
● Conducting exploratory data analysis (EDA) on real-world datasets.
● Cleaning and preprocessing data to prepare for modeling.
2 ● Solving problems related to descriptive statistics (mean, median,
mode, variance).
● Applying probability concepts to data science problems.
● Working with probability distributions and sampling techniques.
3 ● Solving problems involving matrix operations and vector calculus.

4 ● Applying linear algebra concepts to data transformations.

● Implementing supervised models like linear regression, decision trees,
and k-nearest neighbors.
● Implementing unsupervised model like k-means, hierarchical
 ● Implementing anomaly detection for real world data.
● Understanding the concept of overfitting and underfitting through
practical examples.
● Hyperparameter tuning and model evaluation techniques.
6 ● Creating visualizations using Matplotlib and Seaborn.
● Visualizing complex datasets and interpreting the results.
● Building dashboards using tools like Plotly or Dash.

7 ● Implementing a complete machine learning pipeline from data

collection to model deployment.
● Working on real-world datasets and competitions (e.g., Kaggle).
● Understanding the ethical implications and bias in machine learning.

7. Practical Works
S.N. Practical works
1 Conduct an exploratory data analysis (EDA) on a public dataset.
2 Perform data manipulation tasks such as filtering, grouping, and summarizing.
3 Implement and compare different statistical techniques to analyze sample data (e.g.,
hypothesis testing, regression analysis).
4 Clean and preprocess a messy dataset (e.g., handling missing data, encoding
categorical variables, feature scaling).
5 Implement different supervised learning algorithms (e.g., linear regression, decision
trees) on a dataset.
6 Apply clustering techniques (e.g., K-means, hierarchical clustering) on a dataset and
evaluate the clusters.
7 Perform a probabilistic model.
8 Apply anomaly detection methods in real world dataset.

8. Evaluation system and Students’ Responsibilities

Evaluation System
In addition to the formal exam(s) conducted by the Office of the Controller of Examination of
Pokhara University, the internal evaluation of a student may consist of class attendance, class
participation, quizzes, assignments, presentations, written exams, etc. The tabular presentation of
the evaluation system is as follows.

External Evaluation Marks Internal Evaluation Marks

Semester-End 50 Class attendance and participation 5

Examination Lab, Case study and Viva 15
Internal Term Exam 30
Total External 50 Total Internal 50
Full Marks 50+50 = 100
Students’ Responsibilities:
Each student must secure at least 45% marks in the internal evaluation with 80% attendance in the
class to appear in the Semester End Examination. Failing to obtain such a score will be given NOT
QUALIFIED (NQ) and the student will not be eligible to appear in the End-Term examinations.
Students are advised to attend all the classes and complete all the assignments within the specified
time period. If a student does not attend the class(es), it is his/her sole responsibility to cover the
topic(s) taught during the period. If a student fails to attend a formal exam, quiz, test, etc. there
won’t be any provision for a re-exam.

9. Prescribed Books and References

Text Book
Grus, J. Data Science from Scratch: First Principles with Python, Second Edition, O'Reilly
Media.
Geron, A. Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow, Second
Edition, O'Reilly Media.
An Introduction to Statistical Learning by Gareth James et al.
O'Neil, C. Weapons of Math Destruction: How Big Data Increases Inequality and Threatens
Democracy, Crown Publishing Group.
Aggarwal, C. C. (2017). Outlier analysis (2nd ed.). Springer.

Reference Books
McKinney, W. Python for Data Analysis: Data Wrangling with Pandas, NumPy, and IPython,
Second Edition, O'Reilly Media.