Commodity Recommendation System Using Tag Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing commodity recommendation systems fail to adapt to changing user preferences influenced by seasonal, trend, health, and environmental factors, limiting their effectiveness in providing personalized recommendations.
Innovation Solution
A system that detects and ranks tags from user purchase history data, creates feature tables, and combines tags with external environment information to recommend commodities, segmenting users based on attribute and meta-data from various sources like EC sites, SNS, and blogs, to provide dynamic and personalized recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If commodity recommendation is based on purchase history and browsing records, then commodity selection is limited to past behavior patterns, but user preferences change due to season, trend, health condition, and living environment
Solution Approach 1:
The patent segments user preferences into multiple dimensions: basic preference (from purchase history), situational preference (from browsing records), and trend preference (from external information). This segmentation allows the system to handle different aspects of user preference separately and combine them appropriately, resolving the contradiction between relying on past behavior and adapting to changing preferences.
Solution Approach 2:
The system performs preliminary actions by proactively acquiring external environment information (season, trend, health conditions) before making recommendation decisions. This preliminary data collection enables the system to anticipate and adapt to changing user preferences rather than merely reacting to past behavior patterns.
2Measurement precision
If the system integrates multiple data sources (purchase history, browsing records, external environment information), then recommendation accuracy improves, but system complexity increases
Solution Approach 1:
The patent divides the complex preference detection task into separate modules: basic preference extraction from purchase history, situational preference extraction from browsing records, and trend preference extraction from external information. Each module operates independently with its own processing logic, making the overall system more manageable while achieving high precision through comprehensive data integration.
Solution Approach 2:
The patent introduces an intermediary processing layer that standardizes and integrates data from multiple sources before generating recommendations. This intermediary layer acts as a mediator between diverse data sources and the recommendation output, simplifying the system architecture by providing a unified processing framework.
3Ease of manufacture
If the system uses only purchase history data, then data processing is simple, but it cannot capture situational preferences influenced by season, trend, and living environment
Solution Approach 1:
The patent segments preference detection into distinct components: basic preference (purchase history), situational preference (browsing records), and trend preference (external environment information). This segmentation allows the system to maintain simple processing for basic preferences while adding complexity only where necessary for capturing situational and trend preferences.
Solution Approach 2:
The system performs preliminary data acquisition and processing of external environment information (season, trend, health conditions) in advance. This preliminary action prepares the data in a standardized format that can be easily integrated with purchase history and browsing records, making the overall process manageable while capturing comprehensive situational preferences.
Data Source
AI summary
The present invention provides a commodity recommendation system comprising: (1) means for detecting tags from data of purchased commodities based on a purchase history in a user database; (2) means for creating and recording a feature table for each purchased commodity by associating the assigned tags with the purchased commodity; (3) means for ranking the tags with respect to the whole purchased commodity of a user and storing them by repeatedly performing the 1) and the 2) for each purchased commodity; (4) means for registering a list of the ranked tags in the user database in association with the user; (5) means for extracting one or more ranked tags from the user database upon a request and creating a tag combination; (6) means for extracting a commodity that matches the created tag combination; and (7) means for displaying the extracted commodity on a designated terminal.