DV2594 Machine Learning for Streaming Data

Single subject course, 5 Higher education credits, Second cycle, autumn semester 2021

Apply

Apply via universityadmissions.se

The purpose of this course is to give the student an introduction and hands-on approach to the field of machine learning for streaming data.

Facts

  • Type of instruction: Distance, mixed-time, part-time 17%
  • Application code: BTH-D5830
  • Period : 2021-August-30 until 2022-January-16
  • Education level: A1N
  • Application: Apply via universityadmissions.se
  • Language of instruction: The language of instruction is English.
  • Location: Some or all of education and examination is held at distance.
  • No. of occasions: Mandatory: none, Voluntary: none
  • Main field of study: Computer Science
  • Course syllabus: Download
  • Welcome letter: Link to welcome letter from responsible teacher will be posted here no later than 3 weeks before the course begins.
  • Entry requirements: Admission to the course requires a Bachelor of Science in Computer Science, Computer Engineering, Electrical Engineering or similar area.

Content

The course comprises the following themes:
• Understanding and developing machine learning methods and theories including mathematical statistics, dimension reduction, feature/variable selection and visualization, decision trees and its applications, univariate and multivariate linear models, logistic regression, clustering methods, nearest neighbour classifiers, support vector machines, artificial neural networks, ensemble classifiers.
• Discussing application trends of using machine learning methods across industries and in different scientific research problems.
• Designing, implementing, and testing different machine learning algorithms.
• Evaluating machine learning methods using different measurement metrics.

Learning outcomes

Knowledge and understanding
• independently explain and summarize results from the application and evaluation of machine learning methods.
Competence and skills
• implement and apply machine learning methods to different streaming data problems.
• modify existing algorithms or develop new machine learning methods to be applied to different streaming data problems.
Judgement and approach
• plan and execute experiments to evaluate and compare machine learning methods.
• select the best machine learning algorithm by analyzing and evaluating performance of different methods.

Course literature and other teaching material

Course notes will be posted periodically and presentations will be shared weekly on the course webpage.
Main Book: Machine Learning for Data Streams with Practical Examples in MOA
Authors: Albert Bifet, Ricard Gavaldà, Geoff Holmes and Bernhard Pfahringer
ISBN: 9780262037792
Publisher: The MIT Press
Year: 2018
Other teaching materials
Book: Machine Learning (The art and Science of
Algorithms that Make Sense of Data)
Author: Peter Flach
ISBN: 978-1-107-09639-4
Publisher: Cambridge
Year: 2012

Course literature and other teaching material

Course notes will be posted periodically and presentations will be shared weekly on the course webpage.
Main Book: Machine Learning for Data Streams with Practical Examples in MOA
Authors: Albert Bifet, Ricard Gavaldà, Geoff Holmes and Bernhard Pfahringer
ISBN: 9780262037792
Publisher: The MIT Press
Year: 2018
Other teaching materials
Book: Machine Learning (The art and Science of
Algorithms that Make Sense of Data)
Author: Peter Flach
ISBN: 978-1-107-09639-4
Publisher: Cambridge
Year: 2012

Learning methods

The course will be online. The education comprises theory and practical parts. Thus, the course will provide theoretical and practical knowledge to analyze, implement, and evaluate machine learning systems. Moreover, two different assignments and one project will be given during the course and the knowledge is evaluated and increased through assignments and a project. The assignments and project must be conducted individually.

Work placement

No work placement is included in the planned learning activities. BTH is aiming for a close contact with the surrounding community when developing courses and programmes.

Teachers

Course Manager
  1. Hüseyin Kusetogullari

Time allocation

On average, a student should study 134 hours to reach the learning outcomes.
This time includes all the various available learning activities (lectures, self studies, examinations, etc.).
This estimation is based on the fact that one academic year counts as 60 ECTS credits,
corresponding to an average student workload of 1 600 hours. This may vary individually.

Assessments

Grading

The course will be graded G Pass, UX Insufficient, supplementation required, U Fail.

Exams

More information about exams are found in the Student's Portal, where you also enrolls for most exams.


There might be other scheduled examinations. Information regarding these examinations are available in the learning platform Canvas or at other places that the person who is responsible of the course will refer to.

Course Evaluation

The course manager is responsible for the views of students on the course being systematically and regularly gathered and that the results of the evaluations in various forms affect the form and development of the course.

Do you have a question? Type it here and we will get back to you as soon as possible!
×