Keyword Extraction from Patent Claims Using Dependency Structure

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current techniques for patent analysis lack an effective method for extracting keywords from patent documents with high accuracy, despite various approaches proposed in literature.

Innovation Solution

A computer-implemented method that decomposes independent claims into elements, constructs a dependency structure, calculates scores based on element depth, and extracts keywords with scores equal to or above a predetermined threshold, utilizing dependency relations and modifying terms to improve accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional text mining techniques are used for keyword extraction, then the process is simple, but the extraction accuracy is low

Engineering Contradiction:
Improvekeyword extraction accuracyVSAvoidanalysis method complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments patent claims into multiple elements based on grammatical and semantic boundaries. Each element is assigned a depth value in the dependency structure, allowing selective processing of different claim components. This segmentation enables the system to focus computational resources on elements most likely to contain keywords, improving accuracy without uniformly increasing complexity across the entire analysis process.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a depth dimension to the traditional keyword extraction approach by constructing a dependency structure with multiple levels. Terms are evaluated not just based on their presence or frequency, but also based on their positional depth in the claim structure. This additional dimensional information allows the system to distinguish between essential and non-essential terms, significantly improving keyword extraction accuracy.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If dependency structure analysis is implemented, then keyword extraction accuracy improves, but computational complexity increases

Engineering Contradiction:
Improvekeyword extraction accuracyVSAvoidcomputational resources
Core Design Contradiction:
Measurement precisionVSPower

Solution Approach 1:

The patent applies different processing strategies to different elements based on their local characteristics within the claim structure. Elements at different depths in the dependency structure receive different levels of analysis intensity. This localized approach ensures that computational resources are concentrated on elements most likely to yield valuable keywords, rather than uniformly processing the entire claim text.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent performs preliminary decomposition of claims into elements and construction of the dependency structure before the actual keyword extraction process. By pre-organizing the claim structure and identifying element relationships in advance, the system avoids redundant computations during the keyword extraction phase, thereby reducing overall computational complexity while maintaining high accuracy.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS10755049B2Extraction of a keyword in a claim
Publication Date: 2020.08.25 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US10755049B2 patent drawing
  • US10755049B2 patent drawing
  • US10755049B2 patent drawing

AI summary

Extracting one or more keywords in a claim, including: decomposing an independent claim into a plurality of elements, wherein each element in the plurality of elements has at least one term; constructing a dependency structure from the plurality of elements, wherein each element in the plurality of elements has a depth in the dependency structure; calculating a score using the depth of the element corresponding to the at least one term in the independent claim; and extracting at least one keyword from the at least one term having a score equal to or larger than a predetermined threshold.