Browse Prior Art Database

A Method and System for Interactively Recommending Products on an E-commerce Website

IP.com Disclosure Number: IPCOM000237423D
Publication Date: 2014-Jun-18
Document File: 3 page(s) / 83K

Publishing Venue

The IP.com Prior Art Database

Related People

JH Hsiao: INVENTOR

Abstract

A method and system is disclosed for interactively learning a user's product preferences and providing refined product recommendations based on the user's feedback in real time. The method and system allows the user to adjust the recommendation model by providing hints to the system. Based on the hints provided by the user, the method and system immediately refines the model to deliver more relevant products.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

A Method and System for Interactively Recommending Products on an E-commerce Website

Abstract

A method and system is disclosed for interactively learning a user's product preferences and providing refined product recommendations based on the user's feedback in real time.  The method and system allows the user to adjust the recommendation model by providing hints to the system.  Based on the hints provided by the user, the method and system immediately refines the model to deliver more relevant products.

Description

Users generally find it difficult to locate relevant products while shopping on e-commerce websites.  Traditional product recommendation methods in e-commerce websites are based on view-also-view or buy-also-buy models.  The view-also-view (or buy-also-buy) model, which learns from the population behaviors, often just recommends the most popular products to a user, and thus the recommended products hardly satisfy the user's personal product interests.  Moreover, there is no real-time model adaptation based on individual user behavior (e.g., click stream).  Thus, the existing systems fail to capture user’s product intentions during product browsing on e-commerce websites.
Disclosed is a method and system for interactively learning user's product preferences and providing refined product recommendations based on user's feedback in real time.

The method and system provides a two stage recommender, where the initial recommender suggests the products based on population behavior such as, but not limited to, view-also-view and buy-also-buy models.  The second recommender interactively and continuously refines the product recommendations based on user’s feedback and product interests. 

The method and system allows the users to adjust the recommendation model by just giving hints to the system.  Based on the hints provided by the user, the method and system immediately refines the model to deliver more relevant products based on user's feedback.

The following figure illustrates the workflow of the interactive recommender system.

Figure

As illustrated in the figure, the method and system provides an initial list of recommendations to the user based on traditional product recommendation models.  Subsequently, the user draws a shape such as a rectangle around a product to indicate the user’s preference towards that particular product.  Moving on, after acquiring the...