论文部分内容阅读
People today increasingly prefer online over conventional shopping. Online shopping possesses numerous advantages and benefits, such as convenience, lower prices, variety, lack of obligation and discreet purchases. However, selecting from among a huge number of products is a challenge for customers. Customers may spend days or even months viewing relevant products on different web sites, occasionally re-locating previously viewed products for comparison, until a final purchase is made. With the large volume of historically viewed products, it is not an easy job for human users to relocate a previously visited product page using conventional access history lists. Also, the product’s ranking may change, making re-finding it difficult or impossible-even using the same keywords in the same online shopping mall. To address this problem, we developed the ShoppingCat system to assist online buyers in re-finding previously viewed product pages via product-related features or previously accessed context features. We evaluated ShoppingCat’s performance in a 2-month user study: its prediction precision was over 70.0%, and the recall rate was 84.7% particularly for the search-then-browse pages.
People shopping as well as over online shopping. Online shopping possessed huge advantages and benefits, such as convenience, lower prices, variety, lack of obligation and discreet purchases. However, May spend days or even months viewing relevant products on different web sites, occasionally re-locating previously viewed products for comparison, until a final purchase is made. With the large volume of historically viewed products, it is not an easy job for human users to Relocate a previously visited product page using conventional access history lists. Also, the product’s ranking may change, making re-finding it difficult or impossible-even using the same keywords in the same online shopping mall. To address this problem, we developed the ShoppingCat System to assist online buyers in re-finding previously viewed product pages via product-related features or previously accesse d context features. We evaluated ShoppingCat’s performance in a 2-month user study: its prediction precision was was over 70.0%, and the recall rate was 84.7% particular for the search-then-browse pages.