Knowledge Graph for Unstructured Data Entity Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems face challenges in extracting meaningful information from unstructured historical data and handling large volumes of data with different formats, which hinders the efficient determination of equipment operation problems and their solutions.
Innovation Solution
A hardware processor-implemented method that receives operation problem indications, acquires structured and unstructured historical data, determines entity associations based on frequency, and establishes relationships to identify operation solutions, effectively leveraging both structured and unstructured data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If regular association rules and machine learning algorithms are used to analyze historical operation data, then the system can identify relationships between device operations, but it cannot effectively extract hidden information from unstructured data formats
Solution Approach 1:
The patent replaces traditional machine learning algorithms with a knowledge graph-based semantic analysis system. Instead of using statistical methods to find patterns in unstructured data, the system uses entity extraction, relationship identification, and semantic reasoning to interpret the meaning and connections within unstructured text, thereby effectively extracting hidden information that conventional algorithms cannot detect
Solution Approach 2:
The patent introduces a knowledge graph as an intermediary layer between raw unstructured data and analysis results. The knowledge graph stores extracted entities, their attributes, and relationships in a structured semantic format, serving as a bridge that transforms unstructured data into machine-understandable knowledge that can be queried and analyzed effectively
2Reliability
If brute-force approach is used to exhaust different combinations of possible solutions, then all potential solutions can be found, but the solution space becomes impractical to search
Solution Approach 1:
The patent performs preliminary action by building a knowledge graph from historical operation data before actual problem-solving occurs. The system pre-extracts entities, attributes, and relationships from historical records and stores them in the knowledge graph, so that when a new problem arises, the system can quickly query pre-processed knowledge rather than searching through raw data or trying all possible solutions
Solution Approach 2:
The patent segments the solution space by organizing knowledge into discrete entities, attributes, and relationships in the knowledge graph. Instead of treating the solution space as a monolithic search problem, the system divides it into structured knowledge components that can be independently queried and combined, making the problem-solving process more efficient and manageable
3Loss of information
If natural language processing is applied to interpret text descriptions, then some insight can be gained, but it becomes difficult to generate meaning for the whole text when dealing with huge volume of unstructured data in different formats
Solution Approach 1:
The patent segments unstructured text data into discrete entities and attributes using named entity recognition and information extraction techniques. Instead of attempting to process entire text documents as single units, the system identifies and extracts specific entities (e.g., device names, error codes, operations) and their attributes, transforming unstructured text into structured knowledge that can be stored and queried in the knowledge graph
Solution Approach 2:
The patent introduces the knowledge graph as an intermediary that standardizes and structures extracted information from diverse unstructured data sources. The knowledge graph provides a unified schema for storing entities, attributes, and relationships, serving as a mediator that translates various text formats and structures into a consistent knowledge representation that can be efficiently processed and queried
Data Source
AI summary
This disclosure relates generally to data processing, and more particularly, to methods and systems for determining an equipment operation based on historical data. In one embodiment, a hardware processor-implemented method for facilitating an operation of a device is provided. The method comprises: receiving an indication of an operation problem for a first device; acquiring historical operation data of a plurality of devices including the first device, the historical operation data including structured data and unstructured data; determining at least a list of first entities and a list of second entities based on the structured data; determining a set of entity associations, each entity association including at least one of the first entities and at least one of the second entities; determining one or more relationships between each of the entity associations; and determining, based on the one or more determined relationships, an operation solution to solve the operation problem.