Link Prediction Using Semantic and Popularity Score Differences
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing link prediction methods in graph analytics struggle to identify missing relationships between webpages, especially in noisy datasets, and fail to effectively improve user navigation by not accounting for semantic similarity and popularity scores.
Innovation Solution
A system and method that determine similarity and popularity scores between webpages, calculating a difference to identify potential improvements in links by comparing semantic similarity with user interaction data, and recommending link additions or modifications based on a threshold value.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional graph theory methods are used for link prediction, then the system can process network structure data, but it fails to effectively utilize semantic information from webpages
Solution Approach 1:
The patent combines traditional graph theory link prediction with natural language processing (NLP) techniques to create a hybrid system. The NLP component extracts semantic information from webpage content, while the graph theory component analyzes network structure. These two components are merged to generate comprehensive link predictions that consider both semantic similarity and structural relationships, thereby reducing information loss without excessive complexity increase.
Solution Approach 2:
The patent introduces an intermediary NLP processing layer that converts unstructured webpage text into structured semantic features. This intermediary layer bridges the gap between raw semantic information and the graph analytics engine, enabling effective utilization of semantic data while maintaining system modularity and managing complexity through clear separation of concerns.
2Reliability
If link prediction is performed without considering semantic similarity, then the system runs faster, but it cannot identify missing relationships between webpages accurately
Solution Approach 1:
The patent applies preliminary NLP processing to extract semantic features from webpages before the link prediction step. By pre-computing semantic similarity scores and storing them for later use, the system avoids repeated text analysis during link prediction, thereby maintaining high accuracy while reducing processing time during the actual prediction phase.
Solution Approach 2:
The patent segments the link prediction process into distinct stages: semantic feature extraction, similarity scoring, and link prediction. This segmentation allows each component to be optimized independently, with semantic analysis performed once and reuse across multiple predictions, improving both accuracy and efficiency.
3Ease of operation
If the system analyzes all webpage links, then it can provide comprehensive recommendations, but it becomes difficult to manage noisy datasets
Solution Approach 1:
The patent changes the parameter space by transforming raw webpage data into standardized semantic features and similarity scores. This transformation converts noisy, unstructured text data into clean, comparable numerical representations, making the data easier to analyze and reducing the impact of noise while maintaining comprehensive analysis capability.
Solution Approach 2:
The patent replaces manual or simple rule-based link analysis with an automated NLP-driven semantic analysis system. This substitution enables the system to handle noisy datasets more effectively by using sophisticated language understanding algorithms that can filter out noise and extract meaningful semantic relationships automatically.
Data Source
AI summary
A system, non-transitory computer-readable medium, and method are provided. The system comprises at least one processor and memory storing instructions which when executed by the at least one processor configure the at least one processor to perform the method. The non-transitory computer-readable medium has instructions thereon, which when executed by a processor, perform the method. The method comprises determining a similarity score between a first webpage and a second webpage, determining a popularity score of a link between the first webpage and the second webpage, determining a difference between the similarity score and the popularity score, and determining that the link between the first webpage and the second webpage may be improved if the difference is greater than a threshold value.