Entity Search System Using NLP Distance Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current entity recognition systems in unstructured corpuses are inefficient when searching for entities using non-name identifiers, leading to increased processing load and reduced accuracy in identifying relevant mentions.
Innovation Solution
The method employs natural language processing (NLP) to identify candidate entities by analyzing text entries referencing non-name identifiers, calculating distances between entities and identifiers, and selecting the most relevant entity based on these distances, thereby improving the efficiency and accuracy of entity recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional text-based entity search methods are used on unstructured corpuses, then the system can process searches, but the processing load increases and accuracy decreases when searching using non-name identifiers
Solution Approach 1:
The patent introduces an intermediary NLP processing layer between the search query and the corpus. This intermediary extracts and structures entity information from unstructured text, creating a bridge that enables accurate entity identification while maintaining processing efficiency. The NLP component acts as a mediator that transforms raw text into structured entity data that can be efficiently queried.
Solution Approach 2:
The system performs preliminary entity extraction and classification on the corpus before actual search operations. By pre-processing the unstructured text to identify and structure entity mentions, the system prepares the data in advance, reducing the processing load during actual searches and improving both accuracy and efficiency.
2Measurement precision
If comprehensive entity analysis is performed on all text entries, then entity identification accuracy improves, but the time spent searching and reviewing results increases
Solution Approach 1:
The patent extracts only the relevant entity information from the unstructured corpus, separating it from the rest of the text. This extraction process isolates the critical data elements needed for entity identification, allowing the system to achieve high accuracy without processing the entire text corpus during search operations.
Solution Approach 2:
The system segments the unstructured text into discrete entity mentions and attributes. By dividing the corpus into structured entity units, the system can efficiently query and analyze only the relevant segments rather than processing entire text entries, reducing search time while maintaining comprehensive entity identification.
3Productivity
If simple text matching is used for entity search, then the processing load is lower, but the ability to identify entity mentions using non-name identifiers is reduced
Solution Approach 1:
The patent transforms the search parameter from simple text matching to structured entity attribute matching. By changing the search parameter from raw text to structured entity data with specific attributes (such as entity type, relationships, and contextual features), the system maintains processing efficiency while significantly improving the ability to identify entities using various identifier types.
Data Source
AI summary
A computer device receives a request to search a corpus of documents for an entity, wherein the request includes a non-name identifier of the entity. The computing device identifies entries of text within the corpus of documents that reference the non-name identifier. The computing device applies natural language processing (NLP) to content associated with the identified entries within the corpus of documents, wherein the NLP identifies candidate entities associated with the non-name identifier. The computing device selects an entity from the candidate entities based, at least in part, on distances between the candidate entities and references to the non-name identifier in the identified entries. The computing device returns the selected entity to a submitter of the request.