Eva Zangerle

On the impact of text similarity functions on hashtag recommendations in microblogging environments

Microblogging applications such as Twitter are experiencing tremendous success. Microblog users utilize hashtags to categorize posted messages which aim at bringing order to the myriads of microblog messages. However, the percentage of messages incorporating hashtags is small and the used hashtags are very heterogeneous as hashtags may be chosen freely and may consist of any arbitrary combination of characters. This heterogeneity and the lack of use of hashtags lead to significant drawbacks in regards to the search functionality as messages are not categorized in a homogeneous way. In this paper, we present an approach for the recommendation of hashtags suitable for the message the user currently enters which aims at creating a more homogeneous set of hashtags. Furthermore, we present a detailed study on how the similarity measures used for the computation of recommendations influence the final set of recommended hashtags.

Using tag recommendations to homogenize folksonomies in microblogging environments

Microblogging applications such as Twitter are experiencing tremendous success. Twitter users use hashtags to categorize posted messages which aim at bringing order to the chaos of the Twittersphere. However, the percentage of messages including hashtags is very small and the used hashtags are very heterogeneous as hashtags may be chosen freely and may consist of any arbitrary combination of characters. This heterogeneity and the lack of use of hashtags lead to significant drawbacks in regards of the search functionality as messages are not categorized in a homogeneous way. In this paper we present an approach for the recommendation of hashtags suitable for the tweet the user currently enters which aims at creating a more homogeneous set of hashtags. Furthermore, users are encouraged to using hashtags as they are provided with suitable recommendations for hashtags.

Sorry, I was hacked: a classification of compromised twitter accounts

Online social networks like Facebook or Twitter have become powerful information diffusion platforms as they have attracted hundreds of millions of users. The possibility of reaching millions of users within these networks not only attracted standard users, but also cyber-criminals who abuse the networks by spreading spam. This is accomplished by either creating fake accounts, bots, cyborgs or by hacking and compromising accounts. Compromised accounts are subsequently used to spread spam in the name of their legitimate owner. This work sets out to investigate how Twitter users react to having their account hacked and how they deal with compromised accounts. We crawled a data set of tweets in which users state that their account was hacked and subsequently performed a supervised classification of these tweets based on the reaction and behavior of the respective user. We find that 27.30% of the analyzed Twitter users change to a new account once their account was hacked. 50.91% of all users either state that they were hacked or apologize for any unsolicited tweets or direct messages.

#nowplaying Music Dataset: Extracting Listening Behavior from Twitter

The extraction of information from online social networks has become popular in both industry and academia as these data sources allow for innovative applications. However, in the area of music recommender systems and music information retrieval, respective data is hardly exploited. In this paper, we present the #nowplaying dataset, which leverages social media for the creation of a diverse and constantly updated dataset, which describes the music listening behavior of users. For the creation of the dataset, we rely on Twitter, which is frequently facilitated for posting which music the respective user is currently listening to. From such tweets, we extract track and artist information and further metadata. The dataset currently comprises 49 million listening events, 144,011 artists, 1,346,203 tracks and 4,150,615 users which makes it considerably larger than existing datasets.

SnoopyDB: narrowing the gap between structured and unstructured information using recommendations

Knowledge is structured - until it is stored to a wiki-like information system. In this paper we present the multi-user system SnoopyDB, which preserves the structure of knowledge without restricting the type or schema of inserted information. A self-learning schema system and recommendation engine support the user during the process of inserting information. These dynamically calculated recommendations develop an implicit schema, which is used by the majority of stored information. Further recommendation measures enhance the content both semantically and syntactically and motivate the user to insert more information than he intended to.

Guided curation of semistructured data in collaboratively-built knowledge bases

The collaborative curation of semistructured knowledge has become a popular paradigm on the web and also within enterprises. In such knowledge bases a common structure of the stored information is crucial for providing efficient and precise search facilities. However, the task of refining, extending and homogenizing knowledge and its structure is very complex. In this article we present two paradigms for the simplification of this task by providing guidance mechanisms to the user. Both paradigms aim at combining the power of automated extraction algorithms with the semantic awareness of human users to accomplish this refinement task.

The Snoopy Concept: Fighting heterogeneity in semistructured and collaborative information systems by using recommendations

The collaborative creation and manipulation of semistructured data imposes the major problem of structure heterogeneity. The more users enter information, the more heterogeneous the structure of information becomes. This proliferation of the schema has a significantly negative impact on the performance of querying facilities as structured, unified access of data is no longer possible. In this paper we present the Snoopy Concept, a novel approach for collaborative, semistructured information systems within an online environment. It deals with structure heterogeneity by incorporating the user in the alignment process of data already during the insertion. This is accomplished by providing the users with useful recommendations how to structure information. Furthermore, the system encourages users to enter more information as it points users to missing bits of information.

Recommending structure in collaborative semistructured information systems

Semistructured data provides the users of a community-based information system with the flexibility to store information without having to adhere to any predefined, rigid schema. However, such flexibility needs to be used with caution as it can lead to a very heterogeneous data structure and is therefore not feasible in terms of unified data access and search functionality. We present an approach which avoids such proliferation of substructures and provides the inserting user with recommendations, which are responsible for the creation of a commonly used structure. The presented recommendation algorithm adapts the recommendations to the stored information and its structure created by the community.

Improving Context-Aware Music Recommender Systems: Beyond the Pre-filtering Approach

Over the last years, music consumption has changed fundamentally: people switch from private, mostly limited music collections to huge public music collections provided by music streaming platforms. Thus, the amount of available music has increased dramatically and music streaming platforms heavily rely on recommender systems to assist users in discovering music they like. Incorporating the context of users has been shown to improve the quality of recommendations. Previous approaches based on pre-filtering suffered from a split dataset. In this work, we present a context-aware recommender system based on factorization machines that extracts information about the users context from the names of the users playlists. Based on a dataset comprising 15,000 users and 1.8 million tracks we show that our proposed approach outperforms the pre-filtering approach substantially in terms of accuracy of the computed recommendations.

An Empirical Evaluation of Property Recommender Systems for Wikidata and Collaborative Knowledge Bases

The Wikidata platform is a crowdsourced, structured knowledgebase aiming to provide integrated, free and language-agnostic facts which are---amongst others---used by Wikipedias. Users who actively enter, review and revise data on Wikidata are assisted by a property suggesting system which provides users with properties that might also be applicable to a given item. We argue that evaluating and subsequently improving this recommendation mechanism and hence, assisting users, can directly contribute to an even more integrated, consistent and extensive knowledge base serving a huge variety of applications. However, the quality and usefulness of such recommendations has not been evaluated yet. In this work, we provide the first evaluation of different approaches aiming to provide users with property recommendations in the process of curating information on Wikidata. We compare the approach currently facilitated on Wikidata with two state-of-the-art recommendation approaches stemming from the field of RDF recommender systems and collaborative information systems. Further, we also evaluate hybrid recommender systems combining these approaches. Our evaluations show that the current recommendation algorithm works well in regards to recall and precision, reaching a recall@7 of 79.71% and a precision@7 of 27.97%. We also find that generally, incorporating contextual as well as classifying information into the computation of property recommendations can further improve its performance significantly.