Search Resource Recommendation Filtering Boosted Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search modules are unable to effectively differentiate between legitimate and maliciously boosted traffic amounts for application resources, leading to inaccurate recommendations and a poor user experience.
Innovation Solution
A method that analyzes traffic patterns to distinguish between boost-traffic and un-boost-traffic resources by examining download requests and coordinates on a map, adjusting traffic amounts, and determining a recommendation strategy to prioritize un-boost-traffic resources, while potentially shielding boost-traffic resources that are not accurate searches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the search module uses traffic amount as the sole reference for resource recommendation, then resources with high traffic amount are recommended, but maliciously-boosted resources are incorrectly promoted instead of legitimate popular resources
Solution Approach 1:
The patent segments traffic amount into two distinct components: maliciously-boosted traffic amount and legitimate traffic amount. By analyzing download request patterns (coordinates, frequencies, time intervals) and separating these two traffic types, the system can accurately measure only the legitimate traffic component for resource recommendation, thereby resolving the contradiction between measurement accuracy and recommendation reliability
Solution Approach 2:
The patent introduces an intermediary analysis mechanism that examines download request characteristics (coordinates, time intervals, frequencies) to identify and filter out maliciously-boosted traffic. This intermediary layer acts as a mediator between raw traffic data and recommendation decisions, ensuring that only legitimate traffic influences resource ranking
2Device complexity
If the search module recommends resources based on total traffic amount including maliciously-boosted traffic, then the recommendation process is simple, but user experience deteriorates due to irrelevant resource recommendations
Solution Approach 1:
The patent performs preliminary analysis of download request patterns before generating resource recommendations. By pre-identifying and filtering maliciously-boosted traffic through coordinate and time interval analysis, the system prepares clean legitimate traffic data in advance, maintaining a relatively simple recommendation process while ensuring high user experience quality
Solution Approach 2:
The system employs self-service mechanisms where download request data automatically reveals malicious patterns through its own characteristics (unusual coordinate clustering, abnormal time intervals, excessive frequencies). This self-identifying property eliminates the need for complex external verification processes, keeping the recommendation system simple while improving user experience
Data Source
AI summary
A method for search resource recommendation is disclosed. The method includes receiving a search request sent from a terminal, which includes a search keyword; searching a set of resources corresponding to the search keyword and analyzing each resource in the set of resources to determine which resource is a boost-traffic resource and which resource is an un-boost-traffic resource, wherein the boost-traffic resource is a resource having a traffic amount maliciously boosted in a set duration, and the un-boost-traffic resource is a resource having a normal traffic amount; and obtaining a total traffic amount of the un-boost-traffic resource and an adjusted traffic amount of the boost-traffic resource, determining a recommendation strategy of the set of resources according to the total traffic amount of the un-boost-traffic resource and the adjusted traffic amount of the boost-traffic resource, and sending the set of resources to the terminal according to the recommendation strategy.