Deep learning algorithms for recommendation systems

• Κύριος συγγραφέας: Παπαχρονόπουλος, Γεράσιμος
• Άλλοι συγγραφείς: Papachronopoulos, Gerasimos
• Γλώσσα: English
• Έκδοση: 2023
• Θέματα: Deep learning neural networks; Recommendation systems; Cross-domain recommendation systems; Αλγόριθμοι βαθειάς μάθησης; Εισηγητικά συστήματα συστάσεων

Information load is growing daily as a result of the Internet’s, mobile devices’, and e-phenomenal business’s expansion. This results in the creation of a system that can filter and rank the pertinent data for consumers. Recommendation systems have emerged as a solution to this problem, providing consumers with access to individualized knowledge, goods, and services. While researchers have developed various filtering strategies to improve user and system experiences, there is a need to enhance the performance of cross-domain recommendation systems. In this work, we propose a method for multi-domain recommendation systems that leverages multiple datasets from various domains. By training our system on these diverse datasets, we enable it to learn user preferences and behaviors across different domains, resulting in more accurate recommendations. Our method incorporates cross-domain information, allowing the system to transcend traditional domain boundaries and make relevant recommendations. Furthermore, we consider contextual factors such as location, time, and device to further personalize the recommendations. By incorporating these contributions, our method enhances the overall recommendation experience, providing users with a more holistic view and increasing the relevance of recommendations. This advancement in cross-domain recommendation systems has the potential to improve user satisfaction and engagement in an information-rich environment.