Method and apparatus for computing similarity between cross-field documents
US10452696B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Jun 23, 2016 |
| Grant date | Oct 22, 2019 |
| Priority date | — |
| Expiry date | Aug 10, 2037 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06F16/93
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
A method includes storing documents of different fields, and a relationship between any two documents of different fields, performing word segmentation and stop word removal on the documents of different fields, to obtain a vocabulary data set for the documents of different fields, constructing an incidence matrix between the documents of different fields according to the relationship between the any two documents of different fields, obtaining a topic cluster of the documents of different fields according to the vocabulary data set, obtaining a probability that any topic in the topic cluster appears in any document and a matching weight of the any topic for any two different fields according to the incidence matrix and the topic cluster, and computing a similarity between the any two documents according to the probabilities and the matching weight of the any topic for the fields to which the any two documents belong.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.