Use Cases


In projects together with the partners from industry and research, the FAIR-AI research area of the Know-Center has developed several recommender solutions and gained the expertise to create the ScaR recommender framework.


This research framework is a state-of-the art recommendation engine that provides scalable recommendation services nearly of the shelf packed with features that many commercial systems do not have. We used this framework successfully in past and ongoing projects.

Collaboration

If you are interested in setting up a business or research project using the ScaR framework, please contact either Simone Kopeinik, the Deputy Research Area Manager of FAIR-AI, or Dominik Kowald, the Research Area Manager of FAIR-AI.

Simone Kopeinik

Deputy Research Area Manager (FAIR-AI)

Dominik Kowald

Research Area Manager (FAIR-AI)

Job Recommender

The Moshbit GmbH is offering digital services. One is an app called Studo, which simplifies students’ lives and helps them to manage organizational and social aspects of their studies. All the necessary information is integrated in one place and presented to the students in a clear way. On the other hand, Moshbit is developing a job platform called Talto for students and young talents, which should facilitate their career entry. The platform provides career startup information, showcases companies and enables them to find jobs that are of interest to students and young talents, such as bachelor’s theses, part-time jobs or entry-level jobs.
Together with Moshbit, we will research on matchmaking algorithms in the area of student and graduate job markets and explore new ways of automatizing job recommendations through more precise descriptions of the student capabilities, on the one hand, and the offered jobs and their lingua, on the other hand.

Steerable Recommender for Guest Activities

We developed a context-aware recommender which aims to predict a guest's interest in a certain (holiday) activity. After we created guest profiles which include user / age groups, regions, categories, period (last days / week / month). from anonymized Schladming-Dachstein Sommercard data which was provided by the municipality of Schladming, we recommended those activities to guests which they probably care most about.
Beside recommending the most suitable activities to guests, an important aspect was to offer a range of steering opportunties. Hence, the recommeder allows the municipality of Schladming to target guests to certain activities or to (temporarily) exclude activities from results by dynamically taking into account particular activities, whether options, locations and category information for defined time frames and optional boosting settings.

Personalized Event Recommendation

Every year, the Theaterholding Graz with the Graz Opera, the Schauspielhaus, Next Liberty, the Kasematten, the Orpheum and the Dom im Berg registers around 500,000 visitors. As a connection between the cultural-political concerns of the city of Graz and the province of Styria, the expectations and demands of the population as well as the knowledge and skills of the people working in the societies, the theater holding endeavors to generate insights by means of the existing data landscape, which make it possible to develop and strengthen the connection between customers and offers.
Know-Center is providing data and use cases for developing innovative recommender systems for increasing the digital marketing strategy.


Social Book Recommender

In collaboration with our partner Leibnitz-Informationszentrum Wirtschaft (ZBW) we've created a social book recommender system. The recommendations are based on the rental information of two different platforms, which are then combined within the ScaR framework. This enables to create recommendations on a bigger set of interactions and create meaningful recommendations for each platform individually. The main focus has been on the interaction, hence, utilizing collaborative filtering methods for recommending books to user or books for certain bookmark lists. Further in this collaboration we've used the rental interaction data to create a easy to browse and understand bubble visualization.


Project Partner Recommender

OpenAIRE Matchmaker is an innovative new service built upon the OpenAIRE scholarly graph that enables organisations to keyword search for organisations with the strongest records of funding success from the ever-expanding range of funders included in OpenAIRE. Drawing information from OpenAIRE on institutions, projects and funders, OpenAIRE Matchmaker will enable scientific institutions interested in forming consortia to identify potential partners with the exact disciplinary strengths and competences.
Furthermore, we provide personalized recommendations using Collaborative-Filtering and Content-Based Filtering. This recommendation service can be easily embedded into the OpenAIRE portal (or on research funder websites as an OpenAIRE-branded service), resulting in a new OpenAIRE service which demonstrates the added-value potential of the OpenAIRE scholarly graph for researchers and funders.
Findings of the project were accepted at the Open Scholarly Communication workshop of WWW'2018 ( go to publication )

Social Recommender

ScaR provides recommendations to help clients to find the right social counsellors and organizations for their specific problem and help counsellors to answer the clients’ questions by recommending possible answers and related resources. Feedback from a central ticket system and user interactions improve the recommendation mechanism over time and make the system learn what institutions can handle what questions.
Our project partner Nekom offers Know-How and software development in Marketing, Retail and Multichannel E-Commerce. They are also involved in creating an online platform for social organizations to improve the process of serving their clients needs.
Findings of the project were accepted at the RSBDA workshop of i-KNOW'2016 ( go to publication )

Conference Assistent

Conference Assistant is an Android and iOS App developed by the Know-Center with the goal of providing support for the conference organizers as well as conference attendees. Main features of the Conference Assistant App are to provide fast access to the conference schedule, location of the session rooms, information on speakers as well as details on talks and other related events. So far it has been used to support several public conferences such as the World Usability Congress, i-KNOW and Sandbox. ScaR was successfully integrated in the Conference Assistant with the goal of providing personalized recommendations to users. Two types of entities are recommended - talks and people. ScaR utilizes the data from the user and uses traditional recommendation algorithms to match similar users based on their interests or to recommend them talks they might be interested in.
More information: http://demos.know-center.tugraz.at/conference-assistant/

Location Based Recommender

Together with the company Blanc Noir the Know-Center developed a recommender system for the use in both, online and offline marketplaces. While in online marketplaces (e.g., product web shops) a user solely interacts through a virtual environment, in offline marketplaces (e.g., shopping mall) the user physically interacts with the store. The core of the project was to recognize the different activity patterns a user does in a virtual environment (searches through products, liking a Facebook page, etc.) and physical environments (e.g., walks through the shopping mall, dwells in front of a specific store, etc.) and to combine those. In order to achieve the goals, the project has focused on incorporating physical environments (e.g., shopping malls into the loop) by using IPS beacon technology. As such, physical interactions can be tracked in the real world and be leveraged for creating recommendations in the virtual. The other way around also applies, as a user’s online behaviour can mapped and leveraged in the physical world, too.
Findings of the project were accepted at RecSys'2015 ( go to publication )

Brokerage

The Data Market Austria (DMA) offers a platform to bring datasets, data services, consulting, and infrastructure offers to a common marketplace. The recommender systems included in DMA analyses all offerings, to derive suggestions for collaboration between them, like which dataset could be best processed by which data service. The suggestions should help the costumers on DMA to identify new collaborations reaching beyond traditional industry boundaries to get in touch with new clients or suppliers in the digital domain. Human brokers will work together with the recommender system to set up data value chains matching different offers to create a data value chain solving the problems in various domains. In its final expansion stage, DMA is intended to be a central hub for all actors participating in the Austrian data economy, regardless of their industrial and research domain to overcome traditional domain boundaries.
Findings and evaluation results of the project were accepted at the REVEAL workshop of RecSys'2019 ( go to publication )
Ideas on arising new businessmodels coming with such a ecosystem have been published at the New Business Models Conference 2017 in Graz. (.pdf )

Music Recommender

Older people are often faced with dementia. It has been shown, that people with this illness still respond to music they like. Therefore we are developed a music recommender system, which grounds its decisions in biomedical data. All data transfer and functionality is implemented on an android device. In a first step we conducted a study. During this study users had to listen to music while we measured their physiological feedback with a wearable device. After this we developed algorithms that are able to predict the user feedback while listening to a song. In a parallel process we developed a recommender system that link songs with each other (collaborative filtering). In the end we combined all three research and development processes in a mobile application that automatically plays songs to the listener that he or she likes.

Learning Resources

In course of the H2020 project AFEL (Analytics for Everyday Learning), ScaR is used as recommender engine to suggest learning resources to users. Thus, we use concepts of Learning Analytics in order to design a recommender system, which not only accounts for the interests of a user but also for the user's competence level. Our learning resource recommender system will be evaluated in an online study within the Spanish learning platform Didactalia
www.didactalia.net
Next to the Know Center there are 5 other partners involved in the AFEL project beeing GNOSS, Knowledge Media Research Center Tübingen, Leibniz Universitaet Hannover, National University of Ireland Galway and The Open University
Findings of the project were accepted at the SRS workshop of CIKM'2018 ( go to publication )

Hotel Booking

E-Commerce

Together with the fashion retailer MyManou, we developed an E-commerce recommender system. Therefore, we utilized various Collaborative-Filtering and Content-Based Filtering algorithms for the personlized recommendation of fashion articles to users. Apart from that, we introduced filter criteria (e.g., price range, colour, etc.) to enable real-time filtering possibilities for the recommendation results.

Ticket Recommender

Tag Recommender for E-Books

Together with HGV, we developed a tag recommender system to support the e-book annotation process. Therefore, we proposed a hybrid tag recommendation system for e-books, which leverages search query terms from Amazon users and e-book metadata, which is assigned by publishers and editors. In total, we implemented and evaluated 19 algorithms and our results show that we can improve the performance of tag recommender systems for e-books both concerning tag recommendation accuracy, diversity as well as a novel semantic similarity metric.
Findings of the project were accepted at the REVEAL workshop of RecSys'2019 ( go to publication )