Homework

The homework is mandatory, which means that you have to perform some data analysis, summarize its result in a report in 5-10 pages (submit it in PDF) and present your work on the last lecture in 5+5 minutes.

A guide for the selection of homework:

If your MSc/Phd research is related to topics covered at the lecture, then please send me an e-mail (antal at mit.bme.hu) indicating the relevant topic and briefly your intended task (keywords are enough).

Topics with lecturers:

1, Visual data analytics, exploratory data analytics: Prof. András Pataricza/László Gönczy

2, Classification and regression with linear models and classical neural networks (e.g. shallow MLPs): Prof. Tadeusz Dobrowiecki

3, Kernel methods (SVMs, deep neural network/learning, dimensionality reduction, matrix factorization with side information): Bence Bolgár

4, Probabilistic graphical models (Hidden Markov Models, Bayesian networks, factor graphs): Péter Antal

If your MSc/Phd research is not related to these topics or data analysis, then please select a topic (~person), send your preference to me in an e-mail (you can mark multiple topics or order them) and we offer a compact task.

Earlier examples are as follows:

Neural networks: Construction of a neural network and using it for classification   (preliminary English version) Hungarian version can be found here

Matrix factorization: - Landscaping

Probabilistic graphical models: - Learning with Naive Bayesian networks and decision trees

 

 

© 2010-2019 BME MIT