A Cognitively Inspired Clustering Approach for Critique-Based Recommenders
The purpose of recommender systems is to support humans in the purchasing decision-making process. Decision-making is a human activity based on cognitive information. In the field of recommender systems, critiquing has been widely applied as an effective approach for obtaining users' feedback o...
| Autores: | , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión aceptada para publicación |
| Fecha de publicación: | 2020 |
| País: | España |
| Institución: | Varias* (Consorci de Biblioteques Universitáries de Catalunya, Centre de Serveis Científics i Acadèmics de Catalunya) |
| Repositorio: | Recercat. Dipósit de la Recerca de Catalunya |
| OAI Identifier: | oai:recercat.cat:2445/195681 |
| Acceso en línea: | https://hdl.handle.net/2445/195681 |
| Access Level: | acceso abierto |
| Palabra clave: | Sistemes d'ajuda a la decisió Aprenentatge automàtic Algorismes computacionals Decision support systems Machine learning Computer algorithms |
| Sumario: | The purpose of recommender systems is to support humans in the purchasing decision-making process. Decision-making is a human activity based on cognitive information. In the field of recommender systems, critiquing has been widely applied as an effective approach for obtaining users' feedback on recommended products. In the last decade, there have been a large number of proposals in the field of critique-based recommenders. These proposals mainly differ in two aspects: in the source of data and in how it is mined to provide the user with recommendations. To date, no approach has mined data using an adaptive clustering algorithm to increase the recommender's performance. In this paper, we describe how we added a clustering process to a critique-based recommender, thereby adapting the recommendation process and how we defined a cognitive user preference model based on the preferences (i.e., defined by critiques) received by the user. We have developed several proposals based on clustering, whose acronyms are MCP, CUM, CUM-I, and HGR-CUM-I. We compare our proposals with two well-known state-of-the-art approaches: incremental critiquing (IC) and history-guided recommendation (HGR). The results of our experiments showed that using clustering in a critique-based recommender leads to an improvement in their recommendation efficiency, since all the proposals outperform the baseline IC algorithm. Moreover, the performance of the best proposal, HGR-CUM-I, is significantly superior to both the IC and HGR algorithms. Our results indicate that introducing clustering into the critique-based recommender is an appealing option since it enhances overall efficiency, especially with a large data set. |
|---|