Text Mining in Practice (PS, Master)

Text Mining in Practice

Advisors: Dr. Ralf Krestel, Julian Risch

The project seminar's slot is on Wednesdays from 13:30 to 15:00 in F-E.06 on Campus II.

There will be weekly individual group meetings (held in German or English) and occasional meetings with all groups during the seminar slot (held in English) to exchange ideas and present results.

6 ECTS, 4 SWS, ITSE/OSIS/ISAE

Course Description

Text Mining is the process of extracting and deriving information from unstructured text data. It involves Natural Language Processing and Machine Learning Methods, such as Probabilistic Graphical Models and Deep Learning.

This project seminar is about the practical application of Text Mining. Its focus is on a real-world dataset of millions of English-language comments that have been posted by users at discussion sections of online newspapers. Each group of students will work on one out of several different topics, such as:

• Hate Speech Detection
• Comment Volume Prediction
• Comment Ranking
• Modeling User's Commenting Behavior
• Discussion Summarization
• Modeling Linguistic Change

Grading

The grade will consist of

• 20% Mid-Term Presentation
• 20% Final Presentation
• 30% Implementation
• 30% Paper (4 pages, ACM Template)

References

• Machine Learning - The Art and Science of Algorithms that Make Sense of Data by Peter Flach
• Hands-On Machine Learning with Scikit-Learn and TensorFlow by Aurélien Géron
• Deep Learning with Python by François Chollet
• Deep Learning by Ian Goodfellow, Yoshua Bengio, and Aaron Courville

Schedule