Research on digital matching methods integrating user intent and patent technology characteristics
Abstract The hierarchical analysis of user needs, function mapping, and the intelligent extraction and matching of related product patent technical features are crucial for seeking innovative design solutions and improving design efficiency. Patent documents contain rich technical features that prov...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-05-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-03273-4 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract The hierarchical analysis of user needs, function mapping, and the intelligent extraction and matching of related product patent technical features are crucial for seeking innovative design solutions and improving design efficiency. Patent documents contain rich technical features that provide valuable technical references and innovative inspiration for product development. However, the complexity of patent documents and the variability in their textual descriptions make it difficult for designers to utilize patents for technological innovation. To address this issue, this research proposes an integrated method for requirement mining and classification, requirement-function mapping, and intelligent matching of functions with patent technical features, thereby achieving effective mappings between requirements, functions, and technical features. The method consists of four main steps: First, based on the Kano model, this research proposes a G-HOQ method for requirement mining, classification, and function mapping, integrating Grey Relational Analysis (GRA) and the House of Quality (HOQ). This method addresses the issue of reducing decision subjectivity and accurately mining user needs under small sample conditions. Second, based on the characteristics of patent texts, an N-V-N structure (noun-verb-noun) is constructed, and Natural Language Processing (NLP) technology is used to extract patent technical features. Third, the BERT model and Principal Component Analysis (PCA) are applied for vector transformation and dimensionality reduction of feature texts, completing the digital characterization. Fourth, the cosine similarity algorithm is used to identify the required functional needs from patent technical documents and to reorganize solutions based on actual requirements, thereby promoting design innovation. Finally, an innovative design for an earthquake experience platform was used as an example, and four design schemes were developed. The satisfaction of the schemes was evaluated using the fuzzy comprehensive evaluation method, revealing that Scheme A, derived from the proposed method, had the shortest time and the highest overall score of 3.287, thus validating the feasibility and effectiveness of the method. |
|---|---|
| ISSN: | 2045-2322 |