Target Word Recognition Using Multi-Dimensional Characteristic Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current statistics-based methods for target word recognition in Chinese language processing are inadequate due to their reliance on limited characteristic values and linear distribution assumptions, leading to inaccurate results, especially in environments with large volumes of non-linear text data, such as e-commerce websites.
Innovation Solution
A target word recognition technique that uses a computer system to extract characteristic values of any dimension and applies non-linear distribution analysis without human intervention, employing methods like gradient boosting and decision trees to determine screening criteria for accurate recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a single threshold value is used for characteristic values, then the recognition process is simple, but the accuracy of target word recognition deteriorates in non-linear distribution scenarios
Solution Approach 1:
The patent changes from using a single threshold value to using multiple threshold values corresponding to different characteristic values. Each characteristic value has its own threshold, allowing the system to accurately evaluate words with non-linear distributions by comparing multiple parameters rather than a single parameter.
Solution Approach 2:
The patent transitions from one-dimensional single threshold comparison to multi-dimensional threshold comparison by introducing multiple characteristic values (such as word length, character frequency, position frequency, etc.). This dimensional expansion enables comprehensive evaluation of candidate words from multiple angles, improving recognition accuracy for complex non-linear distribution patterns.
2Measurement precision
If multiple characteristic values are used, then the accuracy of target word recognition improves, but the complexity of the recognition system increases
Solution Approach 1:
The patent segments the recognition system into distinct modules: characteristic value extraction module, threshold determination module, and comprehensive evaluation module. Each module handles a specific aspect of the recognition process, making the complex system more manageable and maintainable while achieving high accuracy through multi-parameter evaluation.
Solution Approach 2:
The patent creates a universal recognition framework that can handle multiple types of characteristic values (word length, character frequency, position frequency, semantic features, etc.) through a unified multi-dimensional evaluation mechanism. This multi-functional design allows the system to accommodate various feature types without requiring separate processing paths for each feature.
3Ease of operation
If manual threshold setting is used, then the system is easier to understand, but the productivity and automation level deteriorates
Solution Approach 1:
The patent implements self-service automation where the system automatically extracts characteristic values from text data, determines optimal thresholds using statistical methods or machine learning algorithms, and performs comprehensive evaluation without manual intervention. This automation maintains high understandability through clear algorithmic logic while dramatically improving productivity by eliminating manual threshold setting for each recognition task.
Data Source
AI summary
Target word recognition includes: obtaining a candidate word set and corresponding characteristic computation data, the candidate word set comprising text data, and characteristic computation data being associated with the candidate word set; performing segmentation of the characteristic computation data to generate a plurality of text segments; combining the plurality of text segments to form a text data combination set; determining an intersection of the candidate word set and the text data combination set, the intersection comprising a plurality of text data combinations; determining a plurality of designated characteristic values for the plurality of text data combinations; based at least in part on the plurality of designated characteristic values and according to at least a criterion, recognizing among the plurality of text data combinations target words whose characteristic values fulfill the criterion.