Reinforcement Learning to Personalize User eXperience within Digital Business Ecosystems
Uložené v:
| Názov: | Reinforcement Learning to Personalize User eXperience within Digital Business Ecosystems |
|---|---|
| Autori: | Benramdane, Mustapha Kamal, Kornyshova, Elena |
| Prispievatelia: | BENRAMDANE, Kamal |
| Zdroj: | 2024 IEEE 48th Annual Computers, Software, and Applications Conference (COMPSAC). :584-593 |
| Informácie o vydavateľovi: | IEEE, 2024. |
| Rok vydania: | 2024 |
| Predmety: | Personalization, Digital Business Ecosystem, [INFO.INFO-LG] Computer Science [cs]/Machine Learning [cs.LG], [INFO.INFO-IT] Computer Science [cs]/Information Theory [cs.IT], [INFO] Computer Science [cs], Reinforcement Learning, User eXperience |
| Popis: | Various entities like individuals and organizations, and companies from different market segments that share common interests and business gather in networks and share knowledge, information and services via digital platforms, which lead to the emergence of digital Business Ecosystems (DBE).However, in the ever-evolving landscape of these ecosystems, optimizing and personalizing the User eXperience (UX) becomes a complex task due to the diverse natures and types of data, users' behavior, interactions between different entities, intentions, UX ratings and contextual data. Our concern lies in the presentation of multiple data in a harmonized and heterogeneous format, allowing them to be integrated and exploited as best as possible in processing based on Reinforcement Learning (RL). We first specified a data model detailing the different types of data used related to entities, users, services, and products. Then we dissect a multi-criteria dataset, categorizing it into qualitative and quantitative dimensions. Through rigorous data analysis, we delineate methodologies for data qualification, normalization, and aggregation. In this paper, we provide a detailed analysis of the functioning of this method, the challenges we have faced to proceed multi-dimensional data, which allowed us to provide recommended personalized object via Reinforcement Learning. We proceed with tests on an RL algorithm for UX Object recommendation and discuss the results. |
| Druh dokumentu: | Article Conference object |
| Popis súboru: | application/pdf |
| DOI: | 10.1109/compsac61105.2024.00085 |
| Prístupová URL adresa: | https://hal.science/hal-05100856v1 https://hal.science/hal-04557361v1 https://doi.org/10.1109/compsac61105.2024.00085 |
| Rights: | STM Policy #29 |
| Prístupové číslo: | edsair.doi.dedup.....8b8bb93f6e6dcfe7d6fb0b54d09a19a5 |
| Databáza: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstrakt: | Various entities like individuals and organizations, and companies from different market segments that share common interests and business gather in networks and share knowledge, information and services via digital platforms, which lead to the emergence of digital Business Ecosystems (DBE).However, in the ever-evolving landscape of these ecosystems, optimizing and personalizing the User eXperience (UX) becomes a complex task due to the diverse natures and types of data, users' behavior, interactions between different entities, intentions, UX ratings and contextual data. Our concern lies in the presentation of multiple data in a harmonized and heterogeneous format, allowing them to be integrated and exploited as best as possible in processing based on Reinforcement Learning (RL). We first specified a data model detailing the different types of data used related to entities, users, services, and products. Then we dissect a multi-criteria dataset, categorizing it into qualitative and quantitative dimensions. Through rigorous data analysis, we delineate methodologies for data qualification, normalization, and aggregation. In this paper, we provide a detailed analysis of the functioning of this method, the challenges we have faced to proceed multi-dimensional data, which allowed us to provide recommended personalized object via Reinforcement Learning. We proceed with tests on an RL algorithm for UX Object recommendation and discuss the results. |
|---|---|
| DOI: | 10.1109/compsac61105.2024.00085 |