Information Retrieval System Using Sentence Segmentation for Query Parsing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional information retrieval methods struggle to provide accurate search results for users who are unsure of the keywords for their queries, often returning noisy and irrelevant information, especially when dealing with questioning sentences.
Innovation Solution
An information retrieval system that decomposes and parses sentences into segments and dependency relationships, classifies them as leaf or root segments, and structures the data for efficient searching, allowing for accurate retrieval of missing segments in questioning sentences and generation of answering sentences based on the number of arcs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional keyword-based search is used, then search engine can process queries, but search result varies largely and user cannot reach desired information immediately
Solution Approach 1:
The patent segments sentences into grammatical components (subject, object, verb, etc.) and analyzes dependency relationships between segments. This segmentation enables precise matching of question components with document components, improving search accuracy and reducing time to find desired information.
Solution Approach 2:
The patent introduces an intermediary processing layer that performs morphological analysis and dependency parsing between the user's question and the document database. This intermediary transforms natural language queries into structured segment representations, enabling accurate information retrieval.
2Ease of operation
If user enters a sentence without knowing keywords, then search can be performed, but accurate answer cannot be obtained for questioning sentences
Solution Approach 1:
The system performs self-service by automatically analyzing the grammatical structure and dependency relationships of user input sentences. It identifies question components (such as interrogative words and their corresponding answer components) without requiring user knowledge of keywords, while maintaining high answer accuracy through automated linguistic analysis.
3Productivity
If conventional search returns results, then information is provided, but noise is largely included and results are cyclopaedic
Solution Approach 1:
The patent extracts only the relevant segments from documents that correspond to the question components. By focusing on extracting specific answer segments based on dependency parsing results, it eliminates noise and irrelevant information, providing high-quality search results without unnecessary content.
Data Source
AI summary
The present invention provides an information retrieving apparatus and the like which replies a search result accurately to a question from the user.In the present invention, sentence information of a sentence in collected documents is stored, information of a questioning sentence from the user is received from a terminal 2, the questioning sentence from the user is decomposed into segments (S10), documents having common arc segments are extracted from segments in the questioning sentence from the user, the documents are compared with the questioning sentence, and a leaf segment missing in the questioning sentence is retrieved (S12 to S16), and the search result is transmitted to the terminal 2 (S19).