Automated Product Trend Analysis in User Content
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online systems face challenges in identifying and evaluating product combinations offered by publishing users due to the large volume of content items, making it impractical for publishing users to accurately assess which products are of interest to other users.
Innovation Solution
The online system obtains and analyzes product information from publishing users, applies object detection methods to content items from other users, and uses a machine learning identification model to determine the likelihood of product matches, ranking products based on their inclusion frequency in content items from other users.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the online system obtains a large number of content items including products offered by the publishing user, then the publishing user can identify more product combinations of interest to users, but it becomes impractical for the publishing user to manually identify and evaluate all content items
Solution Approach 1:
The patent introduces an intermediary system (the online system's automated evaluation module) that acts as a mediator between the large volume of content items and the publishing user. This intermediary automatically evaluates content items using machine learning models and object detection, filtering and presenting only relevant information to the publishing user, thus resolving the contradiction between handling large quantities of content and maintaining ease of operation
Solution Approach 2:
The system enables self-service by allowing the publishing user to simply provide product information and images, while the system automatically performs the complex tasks of content item evaluation, product identification, and relevance assessment. The publishing user receives ready-to-use evaluation results without needing to manually process individual content items
2Loss of time
If the publishing user is unable to view all content items provided by certain users, then the publishing user's time and resources are conserved, but the publishing user's ability to accurately evaluate different combinations of products is limited
Solution Approach 1:
The patent replaces the mechanical system of manual content item review with automated computer-based evaluation using machine learning models and object detection algorithms. This substitution maintains measurement precision by systematically analyzing all relevant content items while eliminating the time cost of manual review, as the automated system can process large volumes of content quickly and accurately
Solution Approach 2:
The system implements feedback mechanisms where automated evaluation results are provided back to the publishing user, enabling accurate product combination evaluation without direct user viewing of all content items. The feedback loop ensures that the publishing user receives reliable evaluation data that accurately reflects user interest in product combinations
3Productivity
If the online system uses automated methods to identify products in content items, then the evaluation process becomes more efficient, but the complexity of the system increases
Solution Approach 1:
The patent segments the complex identification process into distinct modular components: object detection modules that identify objects in images, machine learning models that recognize products, and evaluation modules that assess relevance. This segmentation maintains high productivity by allowing parallel processing of different content items while managing system complexity through modular architecture where each component has a specific, well-defined function
Data Source
AI summary
A publishing user identifies a product offered by the user to an online system by providing multiple images of various products viewed at different angles to the online system. The online system applies an identification model to content items obtained from other users to identify one or more of the products in various content items. From a number of content items obtained from other users that include products offered by the publishing user, the online system compute trends. The trends may be for a specific product or products having one or more common attributes. The online system transmits information about the trends to the publishing user or to other users. The trends may also be used to rank recommendations for products to a specific user, where the trends from numbers of content items including products are weighted by the specific user's preferences.