Auto-Complete Suggestion Ranking by Intent Probability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face inefficiencies in searching due to the need to read through numerous auto-completion suggestions for search queries, as existing systems do not effectively prioritize relevant suggestions based on user intent, leading to slowed query progression.
Innovation Solution
A method that receives a partial query, identifies and ranks query suggestions based on their probability of matching user intent, and displays only those with a probability above a threshold, along with indicators to highlight the most relevant suggestions, such as color highlighting, histogram bars, or audible notifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If all query suggestions are displayed equally without ranking, then users can see all possible options, but users must read through numerous suggestions to find relevant ones, slowing down query progression
Solution Approach 1:
The system performs preliminary ranking of query suggestions based on probability calculations before displaying them to the user. By pre-computing and ordering suggestions according to their likelihood of matching user intent, the system eliminates the need for users to manually scan through unranked options, thereby speeding up query progression and reducing time loss.
2Ease of operation
If query suggestions are ranked by probability, then relevant suggestions are prioritized for display, but the system must calculate and compare probabilities for multiple suggestions, increasing processing complexity
Solution Approach 1:
The system employs probabilistic models that automatically calculate and rank query suggestions without requiring manual intervention or complex external processing. The ranking mechanism serves itself by using historical data and statistical patterns to generate probability scores, making the complexity inherent to the task rather than an added burden, while still providing ease of operation for users.
3Loss of time
If only top probability suggestions are displayed above threshold, then users see fewer, more relevant options quickly, but some potentially relevant suggestions below threshold may be missed
Solution Approach 1:
The system applies different display qualities to different suggestions based on their probability scores. Top suggestions above the threshold receive prominent display with visual indicators, while suggestions below the threshold are either not displayed or displayed with reduced prominence. This local differentiation in quality ensures that the most relevant options are immediately visible, reducing time loss, while the threshold mechanism prevents information overload.
Data Source
AI summary
Methods, systems, and apparatuses, including computer programs encoded on a computer storage medium, for optimizing search engine user input selection are disclosed. In one embodiment a user may input a portion of a query term or terms. The method may include receiving a partial query from a user, identifying two or more query suggestions based on the partial query, and determining a probability that each respective complete query suggestion is a query that the user intended to input. The method may also include ranking the two or more query suggestions based on the probability of each respective query suggestion, and determining that a top ranking query suggestion is associated with a probability above a threshold. The method may further include providing for display the two or more query suggestions relative to the partial query, and providing for display, an indication of the top ranking query suggestion among the two or more query suggestions. The user may be able to select an automatically generated search term or terms on-the-fly based on the top ranking query suggestion(s).