Auto-Completion Scoring User Commercial Intent
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current auto-completion systems for text strings entered by users primarily rely on lexicographic matching and previous entries, failing to accurately infer the user's commercial intent behind business name entries, leading to inefficient and irrelevant suggestions.
Innovation Solution
An auto-completion system that calculates a commercial intent score for potential replacement text strings based on user transaction history, business characteristics like category, size, and distance, using TF-IDF for category, and normalized values for size and distance, to prioritize and rank suggestions accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the system uses only lexicographic matching and previous entries for auto-completion, then the system complexity is low, but the relevance and accuracy of suggestions to user commercial intent deteriorates
Solution Approach 1:
The system changes the parameters used for auto-completion from simple lexicographic matching to a multi-dimensional scoring system that incorporates transaction frequency, recency, category matching, and commercial intent signals. This transforms the matching mechanism from a single-parameter approach to a multi-parameter evaluation, improving accuracy while managing complexity through structured computation.
Solution Approach 2:
The patent introduces a commercial intent score as an intermediary metric that mediates between raw transaction data and auto-completion suggestions. This score aggregates multiple factors (transaction history, business characteristics, category relationships) into a single evaluative measure that guides suggestion ranking, simplifying the decision-making process while improving relevance.
2Measurement precision
If the system analyzes transaction history and business characteristics to generate suggestions, then the relevance of auto-completion suggestions improves, but the processing time and computational resources increase
Solution Approach 1:
The system performs preliminary actions by pre-processing and storing transaction history data, business characteristics, and category relationships in structured formats before they are needed for auto-completion. This includes pre-calculating transaction frequencies, organizing business data by category, and establishing normalized value computations, so that during actual auto-completion, the system only needs to retrieve and combine pre-computed values rather than analyzing raw data from scratch.
Solution Approach 2:
The patent applies local quality by computing commercial intent scores and business characteristics only for the specific subset of businesses relevant to the current user and context, rather than processing all possible businesses globally. This localized computation focuses computational resources on the most relevant data points, reducing overall processing time while maintaining suggestion relevance.
3Adaptability or versatility
If the system uses normalized values for category, size, and distance in calculating commercial intent score, then the personalization accuracy improves, but the computational complexity increases
Solution Approach 1:
The system transforms raw business attributes (category, size, distance) into normalized values that are scaled and standardized for comparison. This parameter transformation allows diverse attributes to be combined meaningfully in the commercial intent score calculation, improving personalization accuracy by enabling fair comparison across different attribute types while using consistent mathematical operations.
Data Source
AI summary
Some embodiments provide a system that facilitates use of an auto-completion system. During operation, the system receives an input text string entered by a user. The system determines a set of text strings, wherein each of the text strings in the set can replace the input text string entered by the user. Next, the system calculates a commercial intent score for each text string in the set. Following this, the system selects a replacement text string from the set of text strings based on commercial intent scores computed for each text string in the set. Finally, the system displays the selected replacement text string in the place of the input text string to the user.