Patent Specification Quality Evaluation via Term Frequency Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional technologies are inadequate in evaluating the quality of patent specifications, particularly in determining the fulfillment of support and enablement requirements, which are crucial for acquiring patent rights.

Innovation Solution

An information processing device that separates content from patent documents into unit information pieces, generates frequency distributions, and evaluates the quality by comparing and weighting these distributions to assess the similarity and relevance of terms, thereby providing an accurate support information fulfillment index.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional technologies are used to analyze patent documents, then basic text processing can be performed, but the quality evaluation of specifications including support and enablement requirements cannot be adequately assessed

Engineering Contradiction:
Improvequality evaluation accuracyVSAvoidsupport and enablement requirement fulfillment
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent segmentes the specification into unit information pieces (claims, background, summary, detailed description, drawings) and further divides them into sentences and terms. This segmentation enables precise analysis of each component's contribution to support and enablement requirements, allowing accurate quality evaluation that conventional technologies cannot achieve.

Inventive Principle:
Principle #1Segmentation

2Loss of information

If frequency distribution analysis is performed on patent documents, then term usage patterns can be identified, but the fulfillment of support and enablement requirements cannot be determined

Engineering Contradiction:
Improveinformation completenessVSAvoidrequirement fulfillment assessment
Core Design Contradiction:
Loss of informationVSMeasurement precision

Solution Approach 1:

The patent generates evaluation information that provides feedback on whether support and enablement requirements are fulfilled based on frequency distribution analysis. By comparing the frequency of terms in claims against their frequency in the specification body, the system determines if sufficient support information is provided, thereby assessing requirement fulfillment with precision.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent introduces evaluation information as an intermediary that bridges frequency distribution data and requirement fulfillment assessment. This evaluation information synthesizes the frequency patterns into meaningful indicators of whether the specification adequately supports and enables the claims, solving the gap between raw data and legal requirement assessment.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If patent applications are evaluated without accurate quality assessment, then processing can be completed quickly, but the reliability of patent rights acquisition cannot be ensured

Engineering Contradiction:
Improveevaluation processing speedVSAvoidpatent rights acquisition reliability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent enables automated self-service evaluation of patent specifications by computationally analyzing frequency distributions and generating quality assessments. This automated process maintains high processing speed while ensuring reliability through systematic evaluation of support and enablement requirements, eliminating the need for manual assessment while improving consistency and accuracy.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS10902535B2Information processing device, method and program
Publication Date: 2021.01.26 SHOBAYASHI KAZUKO
  • US10902535B2 patent drawing
  • US10902535B2 patent drawing
  • US10902535B2 patent drawing

AI summary

In this information processing device, a claim term frequency distribution generation unit separates the content of the claims in a publication into individual terms and generates a claim term frequency distribution indicating the frequency distribution of each separated term. A description term frequency distribution generation unit separates the content of the specification in the publication into individual terms and generates a description term frequency distribution indicating the frequency distribution of each separated term. A description synonym frequency distribution generation unit classifies each term extracted from the specification into a plurality of groups respectively corresponding to the plurality of terms in the claims, and generates a description synonym frequency distribution indicating a frequency distribution in which each of the classified plurality of groups serves as a unit. An evaluation information generation unit generates evaluation information on the basis of the claim term frequency distribution and the description term synonym frequency distribution.