Text Analysis Apparatus Using Classification Segmentation for Medical Term Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing natural language processing techniques struggle to accurately extract term expressions and establish correct relationships between terms in medical texts, which are often unstructured and contain descriptions of various organs and diseases.
Innovation Solution
An information processing apparatus that acquires medical texts, classifies attributes of information into fixed units such as sentence, phrase, word, or character units, and performs analysis within each classification using prediction models subjected to machine learning, enabling accurate term extraction and relationship acquisition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If text analysis is performed collectively without classification, then processing efficiency is maintained, but analysis accuracy deteriorates due to mixing different classifications
Solution Approach 1:
The patent divides the text analysis process into multiple segments by classifying text into different categories (e.g., findings, diagnosis, past comparison) before analysis. Each category is analyzed separately using appropriate processing methods, which improves analysis accuracy while managing complexity through structured segmentation.
2Measurement precision
If term extraction is performed on unstructured medical text without classification, then processing speed is maintained, but term extraction accuracy deteriorates
Solution Approach 1:
The patent performs preliminary classification of medical text into structured categories before conducting term extraction. This preliminary action organizes the unstructured text into manageable segments, enabling more accurate term extraction while reducing the overall processing time through efficient structured handling.
3Reliability
If relationships between terms are acquired without classification, then processing simplicity is maintained, but relationship accuracy deteriorates due to medically incorrect relationships
Solution Approach 1:
The patent applies different processing qualities and methods to different parts of the text based on its classification. For example, findings sections are processed differently from diagnosis sections, ensuring that relationship acquisition is contextually appropriate for each medical domain, thereby improving relationship accuracy while managing complexity through localized processing strategies.
Data Source
AI summary
Provided are an information processing apparatus, an information processing method, and a program capable of performing analysis of a text with high accuracy.An information processing apparatus includes one or more processors and one or more memories that store a command executed by the one or more processors. The one or more processors are configured to acquire a text, classify attributes of information described in the text into a fixed unit of the text, analyze the text for each of the same classifications based on a result of the classification, and output a result of the analysis.