Question Utility Scoring via N-gram Probability Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Q&A services cannot accurately determine the most relevant question among those that are equally relevant to a user's query, nor can they differentiate the utility of questions based on their likelihood of being repeatedly asked, leading to suboptimal search results.
Innovation Solution
A question search system that calculates n-gram probabilities and language model utility scores to evaluate the utility of questions, combining these with lexical centrality scores to rank search results, ensuring that the most useful questions are prioritized in search outcomes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Q&A services use traditional vector space model and cosine similarity to rank questions, then questions can be ranked based on keyword closeness, but the system cannot accurately determine the most relevant question among those that are equally relevant
Solution Approach 1:
The patent introduces utility scores as an additional parameter to the existing keyword similarity metric. This new parameter captures the frequency and importance of questions, allowing the system to differentiate between questions that have similar keyword matches but different user utility. The combined ranking mechanism uses both the similarity score and utility score to produce a more accurate relevance determination.
2Ease of operation
If Q&A services treat all questions as plain text and rank them equally based on keyword similarity, then the system is simple to operate, but it cannot determine which of various questions would be most useful to a user
Solution Approach 1:
The system pre-calculates utility scores for all questions in the database before they are needed for search. This preliminary action involves analyzing question frequency, popularity, and other utility metrics in advance, so that when a search is performed, the system can quickly combine these pre-computed scores with real-time similarity matching without adding operational complexity to the user interface.
3Quantity of substance
If the system returns all equally relevant questions, then completeness of results is improved, but the user experience deteriorates due to inability to prioritize most useful questions
Solution Approach 1:
The ranking system dynamically adjusts the order of returned questions based on the combination of similarity scores and utility scores. Questions that are both keyword-matched and high in utility are prioritized higher in the results, while still maintaining a comprehensive set of relevant questions. This dynamic ranking allows the system to balance completeness with user efficiency by making the most useful questions more prominent.
Data Source
AI summary
A question search system provides a collection of questions having words for use in evaluating the utility of the questions based on a language model. The question search system calculates n-gram probabilities for words within the questions of the collection. The n-gram probability of a word for a sequence of n−1 words indicates the probability of that word being next after that sequence in the collection of questions. The n-gram probabilities for the words of the collection represent the language model of the collection. The question search system calculates a language model utility score for each question within a collection that indicates the likelihood that a question is repeatedly asked by users. The question search system derives the language model utility score for a question from the n-gram probabilities of the words within that question.