Andreas Nauerz

Improved recommendation based on collaborative tagging behaviors

Considering the natural tendency of people to follow direct or indirect cues of other peoples activities, collaborative filtering-based recommender systems often predict the utility of an item for a particular user according to previous ratings by other similar users. Consequently, effective searching for the most related neighbors is critical for the success of the recommendations. In recent years, collaborative tagging systems with social bookmarking as their key component from the suite of Web 2.0 technologies allow users to freely bookmark and assign semantic descriptions to various shared resources on the web. While the list of favorite web pages indicates the interests or taste of each user, the assigned tags can further provide useful hints about what a user thinks of the pages. In this paper, we propose a new collaborative filtering approach TBCF (Tag-based Collaborative Filtering) based on the semantic distance among tags assigned by different users to improve the effectiveness of neighbor selection. That is, two users could be considered similar not only if they rated the items similarly, but also if they have similar cognitions over these items. We tested TBCF on real-life datasets, and the experimental results show that our approach has significant improvement against the traditional cosine-based recommendation method while leveraging user input not explicitly targeting the recommendation system.

IntrospectiveViews: an interface for scrutinizing semantic user models

User models are a key component for user-adaptive systems They represent information about users such as interests, expertise, goals, traits, etc This information is used to achieve various adaptation effects, e.g., recommending relevant documents or products To ensure acceptance by users, these models need to be scrutable, i.e., users must be able to view and alter them to understand and if necessary correct the assumptions the system makes about the user However, in most existing systems, this goal is not met In this paper, we introduce IntrospectiveViews, an interface that enables the user to view and edit her user model Furthermore, we present the results of a formative evaluation that show the importance users give in general to different aspects of scrutable user models and also substantiate our claim that IntrospectiveViews is an appropriate realization of an interface to such models.

Ontology-Based Multidimensional Personalization Modeling for the Automatic Generation of Mashups in Next-Generation Portals

Recently, the combination of portal and mashup technology has gained some attention. Portals were originally designed as single points of access to the information and applications distributed across the enterprise. However, due to the increasing number of resources available through portals, they have gained a new challenging goal: To provide users with the information tailored to their individual needs and geared to the situation they are working in. Mashups, the tools that dynamically integrate independent applications, seem to be a good technique to achieve this goal. What is needed, however, are means to automatically create personalized mashups that optimally fit a users information needs in a given situation. In this paper, we describe our approach to this automatic mashup generation. At the core of our approach are different ontology-based models that describe the user, the domain, possible information needs in this domain, and personalization rules determining which information to present to which user in which situation.

An Evolutionary Algorithm for Automatic Composition of Information-gathering Web Services in Mashups

The idea behind mashups is to provide a mechanism that allows for more or less spontaneous combination of existing web applications. Users shall thus be enabled to combine data and services according to their needs.However, existing mashup frameworks require some programming knowledge, hence are not suitable for non-expert users. In this paper, we present a system that builds on existing Semantic Web research to achieve an automatic,ad-hoc generation of mashups thus eliminating the need for programmer involvement. At the core of our approach, there is an evolutionary algorithm that automatically composes different information web services based on semantic service descriptions. The information that has been retrieved from the invoked web services is automatically transformed into a semantic representation and presented as a mashup to the users of the system.

Personalized recommendation of related content based on automatic metadata extraction

In order to efficiently use information, users often need access to additional background information. This additional information might be stored at various places, such as news websites, company directories, geographic information systems, etc. Oftentimes, in order to access these different pieces of information, the user has to launch new browser windows and direct them to appropriate resources. In our todays Web 2.0, the problem of accessing background information becomes even more prominent: Due to the large number of different users contributing, Web 2.0 sites grow quickly and, most often, in a more uncoordinated way regarding, e.g., structure and vocabulary used, than centrally controlled sites. In such an environment, finding relevant information can become a tedious task. In this paper, we propose a framework allowing for automated, user-specific annotation of content in order to enable provisioning of related information. Making use of unstructured data analysis services like UIMA or Calais, we are able to identify certain types of entities like locations, persons, etc. These entities are wrapped into semantic tags that contain machine-readable information about the entity type. The entity types are associated with applications able to provide background information or related content. A location, e.g., could be associated with Google Maps, whereas a person could be associated with the companys employee directory. However, it strongly depends on the individual users interests and experience which additional information he deems relevant. We therefore tailor the information provided based on the User Model, which reflects the users interests and expertise. This allows providing the user with in-place, in-context background information on those entities he is likely to be interested in as well as with recommendations to related content for those entities. It also relieves users from the tedious task of manually collecting relevant additional information. Our main concepts have been prototypically embedded within IBMs WebSphere Portal.

Automating Mashups for Next-Generation Enterprise Portals

The increasing number of resources available through portals establish a need to tailor information to individual needs and situations. Mashups are tools for dynamically integrating independent applications. For portals, what is needed are means to automatically create personalized mashups that optimally fit a users information needs in a given situation. At the core of our approach are different ontology-based models that describe the user, the domain, possible information needs in this domain, and personalization rules determining which information to present to which user in which situation.

Adaptation and Recommendation in Modern Web 2.0 Portals

In this paper, we propose a generic recommender framework that allows transparently integrating different recommender engines into a Portal. The framework comes with a number of preinstalled recommender engines and can be extended by adding further such components. Recommendations are computed by each engine and then transparently merged. This ensures that neither the Portal vendor, nor the Portal operator, nor the user is burdened with choosing an appropriate engine and still high quality recommendations can be made. Furthermore we present means to automatically adapt the Portal system to better suit users needs.

Recommending Background Information and Related Content in Web 2.0 Portals

Modern Web 2.0 Portals have become highly collaborative participation platforms. Users do not only retrieve information, they even contribute content. Due to the large number of different users contributing, Web 2.0 sites grow quickly and, most often, in a more uncoordinated way than centrally controlled sites. Finding relevant information can hence become a tedious task. We will demonstrate a solution allowing for the in-place, in-context recommendation of background information with respect to a certain term or topic and for the recommendation of related content being available in the system. Our solution is based on the extraction of enriched units of information which we either gain automatically via unstructured data analysis or by analyzing user-applied annotations. Our main concepts have been embedded and evaluated within IBMs WebSphere Portal.