Pairwise Item Comparison Data for E-Commerce Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data mining methods fail to effectively assist users in discriminating between alternative items, as they primarily rely on purchase-based relationships, often resulting in poor bundling suggestions and lacking in providing user-preferred item selection insights.
Innovation Solution
The system generates pairwise comparison data by analyzing user activity, identifying which item is more frequently selected over another based on user behavior, and presents this data to users to aid in informed decision-making, incorporating both item viewing and purchasing histories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If purchase-based item relationships are used to generate bundling suggestions, then item relationships can be detected, but the suggestions are poor and do not reflect current user preferences
Solution Approach 1:
The patent changes the parameter basis from purchase-based relationships to view-based relationships. Specifically, it uses the number of times items are viewed together in the same session to generate bundling suggestions, rather than relying solely on purchase history. This parameter change makes the suggestions more reflective of current user preferences and browsing behavior.
Solution Approach 2:
The patent performs preliminary analysis of user viewing behavior to identify items that are frequently viewed together, before users make purchase decisions. By analyzing viewing patterns in advance and generating bundling suggestions based on these patterns, the system can provide more accurate and timely recommendations that reflect current user interests.
2Quantity of substance
If data mining processes identify items viewed in combination, then candidate items can be presented to users, but users cannot effectively discriminate between these candidate items
Solution Approach 1:
The patent segments the large set of candidate items into smaller, more manageable groups based on viewing co-occurrence patterns. By identifying items that are frequently viewed together in the same sessions, the system creates natural segments or bundles of related items, making it easier for users to navigate and discriminate between alternatives without being overwhelmed by the full set of candidates.
3Loss of information
If traditional data mining methods are used, then item relationships can be detected, but user activity data is not fully utilized for generating actionable insights
Solution Approach 1:
The patent makes the data mining system multi-functional by using the same viewing behavior data for multiple purposes: detecting item relationships, generating bundling suggestions, and identifying substitution patterns. This universal approach to data utilization maximizes the value extracted from user activity data, improving both the completeness of information used and the reliability of behavioral relationship detection.
Data Source
AI summary
Data mining systems and methods are disclosed for generating data that is helpful to users in selecting between items represented in an electronic data repository, such as an electronic catalog. One disclosed data mining method generates pairwise comparison data for particular pairs of items. The pairwise comparison data for a given item pair reveals a tendency of users who consider both items in the pair to select one item over the other. The pairwise comparison data may be appropriately exposed to users of the electronic repository. For instance, an item detail page for item A may be supplemented with an indication that users who view both item A and item B select item B a specified percentage of the time. Another data mining method uses item viewing histories and item purchase histories of users in combination to identify pairs of items that are good candidates for being recommended in combination.