Study Guide

Objective

After completing the course the student:
- Can define the main concepts related to machine learning
- Understands the value and the drivers for machine learning
- Can describe the processes of machine learning
- Can use some tools for data analytics and machine learning

Content

Machine learning process and methods
Practical work

Materials

Course book:

Aurélien Géron.
Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow
3rd Edition.
Publisher : O'Reilly Media;
(2022)

We study chapters 1, 2, 3, 4, 6, 9, and 10 of the book. They have about 300 pages, but some are skipped over.

The course book can be read in electronic form from our institution's eBook Central database.

The course also has reading material, which will be announced during the course.

Teaching methods

- Participating in lectures (theory and practice)
- Learning through hands-on programming (classwork assignments)
- Completing homework assignments
- Interacting with the teacher and classmates
- Enhancing knowledge through teamwork projects
- Following the flipped-classroom model (pre-session self-study of theoretical concepts followed by in-class practical application)

Exam schedules

+ No exam, retake not possible after the publication of the final assessment/course grade

International connections

- The course includes approximately 28 (1.5 hours) theory and practice sessions, where students engage with practical tasks.
- Homework exercises will be assigned, with some parts demonstrated during contact sessions.
- Classwork assignments will be given during the practice sessions and must be completed and submitted within those hours to earn points.
- A teamwork project will be introduced in the second month, requiring students to apply their teamwork skills and the knowledge gained from the course to implement their final project.
- A flipped-classroom model may be used for some lectures, where students study the theoretical content at home and focus on practical implementation and discussions during class.

Completion alternatives

The practical work and exercises are primarily conducted using Python, including its DA and ML-related libraries, within Jupyter Notebook.

Student workload

+ Student Responsibilities:
1. Class Participation and doing the classwork assignments:
- The completion of in-class assignments, which must be submitted during class hours.
2. Homework Assignments:
- Completing approximately 9 individual homework assignments, partially demonstrated during contact sessions. The exact number of the assignments will be announced at the first lecture.
3. Final Project:
- A groupwork project (2-3 students) to be completed over weeks 46 & 47, culminating in a presentation in week 48.

+ Student workload:
Contact hours (approximately):
- One introductionary session: 2h
- 13 times 3h theory and practice: 13 x 3h = 39 hours
- Final projects and presentations: 24 hours
- Home work: approximately 75 hours

Total: approximately: 140 hours

Content scheduling

+ The course includes approximately 14 theory sessions and 14 guided practice sessions, 9 personal homework assignments, 8-9 classwork assignment and a teamwork final project

+ Final project is done in groups of 2-3 people outside of guidance sessions. The group sets aside 15 minutes to present the group work during the last session.

+ Content scheduling (Note: this is a preliminary plan and may be adjusted based on class performance.)
- Week 03: Course Introduction (2h)
- Week 04: Landscape of machine learning (3 hours split into separate theory and practice sessions)
- Week 05: Data exploration (3 hours split into separate theory and practice sessions)
- Week 06: Data preparation (3 hours split into separate theory and practice sessions)
- Week 07: Model training, selection, and evaluation (3 hours split into separate theory and practice sessions)
- Week 08: Winter break - Visualization (self-study)
- Week 09: Demonstrations of Exercises 1 – 4 (3h)
- Week 10: Classification (3 hours split into separate theory and practice sessions)
- Week 11: Training models (3 hours split into separate theory and practice sessions)
- Week 12: Decision trees (3 hours split into separate theory and practice sessions)
- Week 13: Unsupervised learning (3 hours split into separate theory and practice sessions)
- Week 14: Guidance to team work (3 hours split into separate theory and practice sessions)
- Week 15: Introduction to Neural networks (3 hours split into separate theory and practice sessions)
- Week 16: Demonstrations of Exercises 5 – 9 (3h)
- Week 17: Team work presentations (3h)

Further information

+ Qualifications/Prerequisites:
Student enrollment in the course will not be accepted by the instructor if they have not passed the following prerequisite courses:
- Python programming skills and skills in utilizing Pandas for data manipulation and Numpy for numerical operations and array handling
- Basic knowledge of probability, statistics and linear algebra

+ Communication Channel:
Itslearning and email

Evaluation scale

H-5

Assessment methods and criteria

2) Students can achieve maximum 200 points from this course that contains:
- Participation and classwork assignments: participating on each lecture and submitting the related classwork assignment during the class hours 4p => 9 X 4 = 36 points.
- Homework assignments: each homework assignment has 10-15 points. There are 9 homework assignments => minimum 90 points and maximum 135 points.
- Teamwork assignment: 29 points

Assessment criteria, fail (0)

The student did NOT get at least 50% of the points in teamwork assignment OR did not get at least 50% of the points in the homework assignments OR did not get at least 50% of the points in participation and classwork assignments.

0-99 points --> Fail

Assessment criteria, satisfactory (1-2)

100-119 points --> 1
120-139 points --> 2

Assessment criteria, good (3-4)

140-159 points --> 3
160-179 points --> 4

Assessment criteria, excellent (5)

180-200 points --> 5