Business Action Extraction via Dependency Structure Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for extracting business actions from daily reports often incorrectly extract phrases unrelated to business actions due to their reliance on natural language analysis, leading to improper generation of information for business activities.
Innovation Solution
An information processing apparatus and method that analyze the dependency structure of sentences in business documents to accurately extract words indicative of business actions by identifying target words and their dependency relations, thereby filtering out irrelevant information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If natural language analysis process is used to extract phrases from text data, then extraction coverage is improved, but extraction precision deteriorates due to inclusion of irrelevant phrases
Solution Approach 1:
The patent segments the extraction process into two distinct stages: first extracting candidate phrases using natural language analysis, then filtering these candidates using a business action dictionary. This segmentation allows the system to maintain comprehensive coverage while achieving high precision by eliminating irrelevant phrases through the dictionary-based filtering mechanism.
Solution Approach 2:
The business action dictionary serves as an intermediary between the natural language analysis process and the final extraction result. It acts as a filter that mediates the output of the NLP process, allowing only phrases that match known business action patterns to pass through, thereby ensuring both comprehensive coverage and high precision.
2Quantity of substance
If comprehensive phrase extraction is performed on all text data, then information coverage is improved, but information quality deteriorates due to inclusion of irrelevant business actions
Solution Approach 1:
The patent extracts only the relevant portion of phrases by using the business action dictionary to identify and extract phrases that correspond to actual business actions. This selective extraction ensures comprehensive coverage of business-related content while maintaining high quality by excluding irrelevant phrases through the dictionary-based filtering mechanism.
3Productivity
If natural language analysis is applied to all text data, then extraction completeness is improved, but extraction accuracy deteriorates due to false positives
Solution Approach 1:
The patent performs preliminary action by first extracting all candidate phrases using natural language analysis, then applies a second filtering step using the business action dictionary. This preliminary extraction followed by targeted filtering ensures that no potential business actions are missed while systematically removing false positives through dictionary-based validation.
Data Source
AI summary
In order to accurately extract a business action from a document such as a daily business report, an information processing apparatus includes: an analysis section that analyzes a dependency structure of a sentence contained in a document concerning a business activity; a target extraction section that extracts, from the sentence, at least one word indicative of a target of a business action; and a business action extraction section that, on the basis of a result of analysis by the analysis section, extracts, as a word indicative of the business action, at least one word having a dependency relation with the at least one word extracted by the target extraction section.