Abstract
Researchers have proposed intelligent product-brokering applications to help facilitate the m-commerce shopping process. However, most algorithms require explicit, user-provided feedback to learn about user preference. In practical applications, users may not be motivated to provide unrewarded and time-consuming feedback. By adopting a cognitive approach, this paper investigates the possibility of replacing user feedback with user behavioral data analysis during product browsing. By means of evolutionary algorithms, the system is able to derive corresponding models that simulate the user's shopping behavior. User group profiling is also implemented to help identify the user's shopping patterns. Upon simulations of trial cases with consistent and rational shopping patterns, our experimental results confirm this approach being promising. The system shows high accuracy in detecting the preferences of the user. The algorithms are also portable and effective across different products.
Original language | English |
---|---|
Pages (from-to) | 401-427 |
Number of pages | 27 |
Journal | International Journal of Computational Intelligence and Applications |
Volume | 7 |
Issue number | 4 |
DOIs | |
Publication status | Published - 2008 |
Keywords
- Genetic algorithms
- Product brokering
- User feedback
- User preference