Text Classification via Semantic Unit Keyword Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current text classification methods, whether manual or rule-based, are inefficient, lack accuracy, and require extensive labeled training data, making them costly and inflexible, especially in automated testing where large volumes of test logs need to be analyzed.
Innovation Solution
The method employs AI technology by converting text semantic units into text keywords using a mapping table, extracting semantic features with a semantic model, and classifying using a classification model, reducing the need for labeled data and improving accuracy through unsupervised learning and semantic clustering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual text processing is used, then flexibility and adaptability are maintained, but efficiency and productivity are low
Solution Approach 1:
The system performs automatic text classification through semantic unit extraction and keyword mapping without requiring manual intervention for each text item. The classification model automatically processes text data by extracting semantic units, mapping them to keywords, and generating classification results, enabling the system to serve itself rather than requiring continuous manual operation.
Solution Approach 2:
The patent replaces manual mechanical text processing with an automated computational system. The semantic model and classification model substitute human cognitive processes with machine learning algorithms, automatically extracting semantic units, mapping them to keywords through predefined correspondence relationships, and performing classification without manual intervention.
2Adaptability or versatility
If rule-based text processing is used, then consistency and reliability are improved, but flexibility and adaptability are reduced
Solution Approach 1:
The system changes the parameter of text representation from raw text to extracted semantic units and mapped keywords. By transforming text into a standardized format of semantic units that correspond to predefined keywords, the system maintains consistency in processing while adapting to different text inputs. The classification model then operates on these standardized parameters rather than raw text.
Solution Approach 2:
The patent introduces semantic units and keywords as intermediary representations between raw text and classification results. Instead of directly classifying raw text, the system first extracts semantic units, maps them to keywords through correspondence relationships, and then performs classification on the keyword representation. This intermediary layer provides both flexibility in handling diverse text inputs and consistency in the classification process.
3Measurement precision
If AI-based text classification is used, then accuracy is improved, but the need for labeled training data and model complexity increase
Solution Approach 1:
The patent segments the text classification task into distinct components: semantic unit extraction, keyword mapping, and classification. Instead of using a single complex end-to-end model, the system divides the process into manageable stages where semantic units are extracted from text, mapped to predefined keywords, and then classified. This segmentation reduces the complexity of individual components while maintaining overall accuracy.
Solution Approach 2:
The system extracts semantic units from raw text as a separate intermediate step before classification. By taking out the semantic extraction process and handling it separately through predefined semantic unit patterns and keyword mappings, the system reduces the burden on the classification model itself. The classification model only needs to work with the extracted semantic keywords rather than processing raw text directly, simplifying its requirements.
4Measurement precision
If extensive labeled training data is used, then classification accuracy is improved, but cost and time consumption increase
Solution Approach 1:
The patent performs preliminary action by pre-defining correspondence relationships between semantic units and keywords before the actual classification task. Instead of requiring extensive labeled training data during the classification phase, the system establishes semantic unit patterns and their corresponding keywords in advance. This preliminary preparation creates a reusable mapping framework that can be applied to new text data without requiring additional labeled examples for each classification task.
Data Source
AI summary
Disclosed is a text classification method and a text classification device. The text classification method includes: receiving text data (S1), the text data comprising one or more text semantic units; replacing the text semantic unit with a corresponding text keyword (S2), based on a correspondence between text semantic elements and text keywords; extracting, with a semantic model, a semantic feature of the text keyword (S3); and classifying, with a classification model, the text keyword at least based on the semantic feature, as a classification result of the text data (S4).