Our aim is to design, develop and evaluate novel recommender systems that are based on theories and models of human behaviour. We study potentially harmful effects of these systems such as filter bubbles and biases that impact users of recommender systems. We offer access to boost the KPIs of your online platform (e.g., engagement or revenue) with personalized recommendations.
SERVICE DESCRIPTION
Our aim is to design, develop and evaluate novel recommender systems that are based on theories and models of human behaviour. We study potentially harmful effects of these systems such as filter bubbles and biases that impact users of recommender systems. We offer access to boost the KPIs of your online platform (e.g., engagement or revenue) with personalized recommendations.
To do that we help with:
- Exploring and analysing the data of online user behaviour to understand the preference of your customers and find potentially hidden groups of users which may have been neglected
- Simulating the customer behaviour in an offline setting to find out what algorithms are best suited to generate personalized content recommendations
- Evaluating on different State-of-the-Art methods to find out the true utility of algorithms not only with respect to accuracy, but also different biases
- Adapt our production ready recommender system for the respective domain-dependent data requirements
- Integrate and evaluate in an online setting the true utility of the developed recommendation algorithms
SPECIAL ACCESS CONDITIONS
No
