Patent Matching Analysis System Using NLP Feature Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing patent search and analysis processes are inefficient due to the difficulty in selecting proper keywords and identifying effective search strategies, leading to the need for specialized firms and cumbersome manual processes to find relevant patents and court documents.
Innovation Solution
A patent matching analysis system that utilizes natural language processing (NLP) to parse textual information from patents and court documents, extracting features and transforming them into vectors to identify matching and related patents, and generating ontology reports for court and patent cases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual patent search and analysis processes are used, then users can obtain patent information, but the process is time-consuming and requires specialized knowledge
Solution Approach 1:
The patent replaces manual mechanical search processes with an automated computing system that uses NLP and machine learning algorithms to parse patent documents, extract features, and identify matching patents. The system automatically retrieves patent documents from data systems, processes textual information through NLP engines, and generates matching results without requiring manual intervention in the search process.
Solution Approach 2:
The system enables self-service patent analysis by providing automated tools that allow users to perform complex patent searches and analysis without needing specialized patent research expertise. The computing system handles document retrieval, parsing, feature extraction, and matching independently, making the process accessible to users without specialized training.
2Ease of operation
If specialized patent research firms are hired, then accurate patent searches can be performed, but the cost and complexity increase
Solution Approach 1:
The computing system performs multiple functions within a single integrated platform: retrieving patent documents from various data systems, parsing textual information through NLP, extracting relevant features, identifying matching patents, and generating comprehensive reports. This multi-functional system replaces the need for specialized external firms while maintaining ease of operation through a unified interface.
3Loss of information
If comprehensive patent analysis is performed manually, then detailed ontology information can be obtained, but the process becomes cumbersome and inefficient
Solution Approach 1:
The system continuously processes patent documents through an automated workflow that retrieves documents, parses textual information, extracts features, and generates reports without interruption. The NLP engine continuously analyzes patent texts to extract ontology information, and the system maintains continuous operation to process large volumes of patent data efficiently, eliminating the stop-and-start nature of manual analysis.
Data Source
AI summary
A patent matching analysis system receives an input indicating a source patent or a court case. When the input indicates a court case, the system identifies a patent associated with the court case, which is deemed as a source patent. For each source patent, the system retrieves a source patent document, parses textual information of the source patent document using an NLP engine, extracts a first set of features, and generates a first feature vector. The system then identifies multiple candidate patents. For each of the candidate patents, the system retrieves a candidate patent document, parses the textual information, extracts a second set of features, and generates a second feature vector. The system then determines a similarity between the first and second feature vectors. Based on the determined similarities, the system identifies one or more target patents, and visualizes the source patent and the determined target patents.