Machine Learning Patent Classification Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technical classification codes for patents, such as IPC and CPC, often fail to accurately match technologies disclosed in patent documents, leading to inefficiencies in classification, increased time and cost, and difficulties in managing patent durations for patent owners.
Innovation Solution
A method using machine learning to learn classification patterns from patent documents, generating a classification model based on similarity, and predicting user-defined classification standards to automate the classification and duration management of patent documents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If technical classification codes (IPC, CPC) are used for patent classification, then classification can be performed using standardized codes, but the classification accuracy and matching degree with actual technologies deteriorate
Solution Approach 1:
The patent introduces an intermediary classification model trained on user-specific classification patterns that mediates between the standardized IPC/CPC codes and the actual technology content. This intermediary layer learns the mapping relationships and transforms standardized codes into user-relevant classifications, resolving the conflict between standardization and accuracy.
Solution Approach 2:
The system changes the classification parameters by training machine learning models on user-specific classification data, transforming the rigid standardized classification parameters into adaptive, user-customized parameters that better match actual patent technologies while maintaining the structured nature of classification codes.
2Measurement precision
If users directly identify and classify all patent documents, then classification accuracy improves, but time consumption and cost increase
Solution Approach 1:
The system enables self-service classification by training machine learning models on user classification patterns, allowing the system to automatically perform classifications in the user's style without requiring continuous manual intervention. The model learns from user corrections and improvements, progressively reducing the time users need to spend on classification while maintaining high accuracy.
Solution Approach 2:
The system performs preliminary classification using the trained model before user review, pre-processing the classification work in advance. This preliminary action handles the majority of classifications automatically, reducing the time users need to spend on manual identification while maintaining accuracy through subsequent user feedback and model refinement.
3Measurement precision
If users manually calculate and manage patent duration for each technology, then duration management accuracy improves, but management complexity and time cost increase
Solution Approach 1:
The patent replaces the manual mechanical process of calculating and tracking patent durations with an automated computational system. The machine learning model automatically calculates duration based on classification results and patent data, eliminating the need for manual calculation while maintaining or improving accuracy. This substitution reduces management complexity by automating the entire duration management workflow.
Data Source
AI summary
A method for classifying patent documents may include the steps of identifying original patent documents in a patent pool identified according to a search and/or selection input of a user, determining the first patent documents classified in the classification generated by the user, among the original patent documents in the patent pool; outputting an indication for accuracy related to the automatic classification; and performing the automatic classification for second patent document which are not yet classified, among the patent documents in the patent pool.