In November 2017 a research cooperation with adidas was started. The main goal of this project is to understand user navigation and browsing behaviour in web shops and to develop and improve algorithms to personalize the online experience of users.
More and more people shop online and as the market grows it is now more important than ever to have a user-friendly web shop. Helping users to find interesting things through the use of useful and accurate recommendations is one way to improve the experience for users and to increase loyalty. But also if the goal is not actually recommending a product, but just showing interesting content, it is of great importance to know what a user likes and what not. Thus, the first step is to get to know the user and his likings.
For the analysis of user behaviour in web shops there are several sources to be taken into account, like user profiles, browsing behaviour and also sales or subscription to newsletters. Personalization based only on browsing behaviour is also be of special interest, as many users of web shops don’t have an account.
Building upon the results of the analysis there will be several possibilities to apply and develop methods for personalization like page or product recommendation or support navigation and also connecting users with similar interests.
One great challenge but also opportunity of the project is the availability of Big Data for analysis and development.
Further the evaluation of the used and developed methods are part of the project. How the provided support is perceived by the user is also an important finding, apart from using numerical metrics to measure the capabilities of the system.
1.Fischer, E., Zoller, D., Hotho, A.: Comparison of Transformer-Based Sequential Product Recommendation Models for the Coveo Data Challenge. SIGIR Workshop On eCommerce. (2021).
2.Fischer, E., Zoller, D., Dallmann, A., Hotho, A.: Integrating Keywords into BERT4Rec for Sequential Recommendation. KI 2020: Advances in Artificial Intelligence (2020).