Department of Computational Perception
Department of
Computational Perception
 Johannes Kepler Universität Linz

Home  –  Mission  –  Teaching  –  People  –  Research  –  Media  –  Awards  –  Impressum

WEB & SOCIAL MEDIA

<< back to overview


We are actively pursuing research related to Web and Social Media Analysis and Mining, e.g., extraction and retrieval of music information from Web pages, blogs, or platforms such as YouTube, Facebook, or Twitter. Many of the following topics are strongly connected to music and multimedia applications. However, this is just one possible application scenario. We are also open for general Web mining projects, such as Web crawling, information retrieval, and extraction, and more specific applications such as intelligent information management, personalized recommendation, user-behavior analysis, and opinion mining.


Music Information Extraction and Music Information Systems

The Web of music contains a wealth of different types of information on music artists, albums, songs, genres, etc., scattered over various platforms and services. Examples are biographies, band member histories, country of origin, discographies, genres, tags, song lyrics, and album covers. Besides specific platforms and services dedicated to music, such as last.fm, allmusic.com, or the echonest, pieces of relevant information can be found on fan pages and blogs. Music Information Extraction aims at identifying and extracting these pieces of information from unstructured and semi-structured Web sources. A possible goal is to combine the various types of extracted meta-data and relations into a Music Information System.

Examples for projects:
  • Automatic detection of a person's or institution's nationality
  • Extracting and visualizing semantic relations (e.g., between music artists or songs)
  • Detecting cover versions of songs
  • Music-related image retrieval (band pictures, album covers, logos, etc.)
  • Information extraction from videos (e.g., YouTube)
  • Novelty detection of bands, album releases, tours, etc.
  • Personalized music news-system
  • Visual Web Page Segmentation
Contact: Markus Schedl, Peter Knees

Music Recommendation and Discovery

Using the Web, one can find many potential sources for discovery and recommendation of similar artists or songs, for instance, word patterns on music pages, user-generated tags, ratings, tweets, search engine page counts, peer-to-peer networks, or playlists. Mining these sources can not only give information on the artistic context of music and help estimating similarity, but also to defer measures and indicators such as popularity, hotness, novelty, and genre prototypicality.


Examples for projects:
  • Implementation and comparison of different context-based similarity measures
  • Experimenting with different and adaptive query settings for page retrieval
  • A Web crawler focused on music related pages
  • Detecting popular music artists and songs
  • Predicting up-and-coming music artists and songs
  • Visualizing "Popularity Flows" around the world using Twitter or peer-to-peer network data
Contact: Peter Knees, Markus Schedl

Social Media Mining and Social Network Analysis

The Social Web enables its users to build virtual networks and to comment on, interact with, rate, and contribute content of all types. Thus, social networks, (micro-)blogs, image-, video-, and music-sharing platforms, or RSS-feeds provide a neverending stream of harvestable knowledge. However, finding the gems is non-trivial. Analyzing link and network structures, exploiting the "wisdom of the crowd", processing the plethora of user-generated multimedia content, and filtering irrelevant material and noise can help the user in dealing with this load of data and satisfying her information need more efficiently.

Examples for projects:
  • Recommending users and content from friendship relations
  • Deriving music similarity from Social Media sources
  • Learning location from tweets' content
  • Localization of flickr images based on image and tag similarity
  • Age, gender, and personality prediction from user-generated content
  • YouTube for multimedia information retrieval (analyzing tags, comments, playlists, image previews, related videos, recommendations, etc.)
  • Popularity estimation from social sources (e.g., Twitter, Facebook, IMs, etc.)
Contact: Markus Schedl, Peter Knees

Personalized Music Retrieval and Recommendation

One of the biggest challenges for today's users of digital music collections is how to find music that matches his or her taste depending on factors like the music content, the user's general music taste, but also his or her current emotional state, activity, situation, or surrounding. The objective is to model such aspects, i.e., the user context, via exploiting information collectable during listening, for example on today's smart phones. This includes location, temperature, movements, used playback device, and listening context. Possible applications are systems for personalized access to music collections, user-adaptive playlist generation, or adaptive user interfaces to access music collections.

Examples for projects:
  • User profiling and modeling (e.g., for recommendation and similarity estimation)
  • Context-aware (e.g., Geolocation-aware) services
  • Assessing the potential of different data sources to model user profiles (possible application in the music domain)
  • Sensor-based activity detection on Android smartphones
  • A Web survey about mining user context
  • Privacy issues in personalized systems
  • Learning and recommendation from Web bookmarks
  • Personalized publication library and visualization (similar to the ISMIR Cloud)
Contact: Markus Schedl

Opinion Mining and Sentiment Analysis

"Sentiment analysis or opinion mining refers to the application of natural language processing, computational linguistics, and text analytics to identify and extract subjective information in source materials. Generally speaking, sentiment analysis aims to determine the attitude of a speaker or a writer with respect to some topic or the overall contextual polarity of a document. The attitude may be his or her judgement or evaluation (see appraisal theory), affective state (that is to say, the emotional state of the author when writing), or the intended emotional communication (that is to say, the emotional effect the author wishes to have on the reader)." [Wikipedia]

Examples for projects:
  • Comment and posting analysis
  • Automatic rating prediction
  • Controversy detection
  • Diversity mining
  • Spam-comment filtering
Contact: Markus Schedl, Peter Knees



last edited by pk on 2012-02-20