DM2583 Big Data in Media Technology 7.5 credits

Big data i medieteknik

Digital media information is streaming in from all sorts of sensors, instruments, and simulations, overwhelming our capacity to organize, analyze, and store it. We need powerful new tools to analyze, mine, visualize,  and manipulate big-media data and must rethink our whole approach to data-intensive science as the fourth scientific paradigm. In the new paradigm, we mine big data, looking for relationships and correlations, using software toolkits and programming methods to discover interesting patters and possible rules that govern them. Big data, as a data science is one of the most important sciences for the media industry and its importance is expected to continue to grow for the foreseeable future.

  • Education cycle

    Second cycle
  • Main field of study

    Computer Science and Engineering
  • Grading scale

    A, B, C, D, E, FX, F

Course offerings

Spring 19 for programme students

Intended learning outcomes

After the course, the students should be able to

• account for basic methods, technologies and tools in big data analysis,

• give examples of how big data analysis can be applied on our permanently growing betray of data by searching interesting patterns in data, and how one makes them foreseeable through modelling and visualization techniques,

• explain how one can study the consumers' media use and purchase behaviour with big data analysis,

• account for recommendations system engineering in relation to the customers' trust, knowledge, loyalty and other social aspects of them,

• create value in the media branch of trade through big data analysis,

• use scientific big data technologies, tools and methods to solve practical problems in media technology,

• design the most important stages in big data work from collecting, preparing and modelling data to evaluation and dissemination of results

• explain important machine learning concepts such as feature extraction, cross validation, generalisation, over fitting, prediction and the curse of dimensionality,

• account for how commonly occurring data modelling methods work, what are their applications, and describe their assumptions and limitations,

• apply common data modelling frameworks, technologies and tools within a broad spectrum of media application areas,

• apply and evaluate results that derive from use of common data modelling frameworks, by means of Matlab or Python,

• use tools and visualization techniques to evaluate models, identify patterns and data functions.

Course main content

• Basic methods: pattern recognition, machine learning, data analysis, data visualization, network analysis, and commonly used toolboxes for data mining and visualization of big data.

• Case studies (seminar): students study selected cases and use course methods for analysis and data visualization.

• Small student projects: the students use big data methods for, e.g., studying consumer media use and purchase behaviours. Will be presented as a short research report.

Eligibility

Degree of Bachelor or the equivalent. SF1604 Linear Algebra, SF1625 One variable calculus, SF1626 Multivariable analysis, SF1901 Probability and Statistics or the equivalent. Basic knowledge in Matlab or Python.

Literature

Examination

  • LIT1 - Literature, 2.5, grading scale: P, F
  • PRO1 - Project, 5.0, grading scale: A, B, C, D, E, FX, F

Offered by

EECS/Human Centered Technology

Examiner

Haibo Li <haiboli@kth.se>

Version

Course syllabus valid from: Autumn 2017.
Examination information valid from: Spring 2019.