Existing systems for recommending hotels to mobile travelers are subject to several problems. For example, a traveler might choose a dominated hotel that is inferior to another hotel in all aspects. This problem cannot be solved by simply changing the weights assigned to the attributes of a hotel. In addition, a nonlinear recommendation mechanism, instead of a linear one, may be more effective for tailoring the recommendation result to a traveler’s choice. To address these concerns, this study applied two treatments. First, an artificial attribute is added to each hotel to model a traveler’s unknown preference for that hotel. The value of a traveler’s unknown preference is determined by solving an integer nonlinear programming problem. Subsequently, a backward propagation network is constructed to map the recommendation results to travelers’ choices, to improve the successful recommendation rate. The effectiveness of the proposed methodology was evaluated in a field study conducted in a small region of Seatwen District, Taichung City, Taiwan, and the experimental results supported its superiority over several existing methods in improving the successful recommendation rate.
|Number of pages||8|
|Journal||Journal of Ambient Intelligence and Humanized Computing|
|State||Published - 1 Apr 2018|
- Backward propagation network
- Integer nonlinear programming