Topics for Projects and Thesis @IWM/KC

Sehr geehrte Studierende,

Ich möchte Sie auf 2 aktuelle Diplomarbeitsausschreibungen am Institut für Wissensmanagement hinweisen. Beide Themen eröffnen die Möglichkeit zum Erhalt eines Stipendiums im Rahmen eines vom FWF geförderten Forschungsprojekts.

1) Intent Analysis: Mining and Correlating Human Goals from Social Media Corpora
2) Simulation of Cooperation and Diffusion in Social Media

Nähere Details finden Sie unter: http://kmi.tugraz.at/staff/markus/open-topics/

Sollten Sie Interesse und/oder Fragen haben, melden Sie sich bitte via e-mail oder telefonisch zur Vereinbarung eines unverbindl. Informationsgesprächs - Ich stehe gerne für weitere Auskünfte zur Verfügung.

mit vielen Grüssen,

Markus Strohmaier

You will implement a Firefox Plugin that allows participants of a study to document their goals behind each 1) search query entered 2) each URL typed into the address bar. You will implement the Plugin and corresponding logging data, design and execute a user study and evaluate the results. This is a larger topic suitable for a follow-up master’s thesis.

Your task is to develop and evaluate goal/purpose oriented extensions for bibsonomy.org. This is a larger topic suitable for a follow-up master’s thesis. Related Work

Based on existing goal-oriented modeling languages such as i*, your task is to develop a web-based representation format for multi-media and textual resources. Related Work

Using existing goal taxonomies, your task is to create intent profiles (what goals is a given text about) of textual resources on the web (e.g. blogs).

Using Formal Concept Analysis, you will study the deduction of hierarchical structures from folksonomies. (Some) datasets will be provided (e.g. 43things.com). Related Work

Folksonomies represent an omnipresent concept on the “web 2.0″. They typically consist of tri-partite graphs (Object - User - Tag) that are constructed by large sets of users (e.g. del.icio.us, 43things.com, etc). Using MatLab and existing MatLab toolboxes, you will analyse provided n-partite graphs from an information-theoretic (e.g. Entropy) and a graph-theoretic (e.g. Tensors) perspective. This is a larger topic suitable for a follow-up master’s thesis. Related Work

Using the goal-oriented modeling language i*, you will develop a method to provide goal-oriented access to a website’s resources. The project consists of the following tasks: Chosing a website, Identifying and modeling user goals, representing user goals, implementing a prototype, evaluating the solution. Related Work

The institute has access to two datasets containing several million queries from AOL/Microsoft (~20 Mio + ~ 15 Mio queries). Applying natural language processing and machine learning methods, your task is to study the nature of goals in search query logs.

Using existing goal modeling and goal taxonomies, your task is to develop a new type of social software that allows users to maintain a list of goals and corresponding textual resources. Integration with existing social platforms (e.g. Facebook, Twitter, etc) is possible. This is a larger topic suitable for a follow-up master’s thesis.