An integrated solution for detecting rising technology stars in co-inventor networks

Lin Zhu, Donghua Zhu, Xuefeng Wang, Scott W. Cunningham, Zhinan Wang

Research output: Contribution to journalArticle

Abstract

Online patent databases are powerful resources for tech mining and social network analysis and, especially, identifying rising technology stars in co-inventor networks. However, it’s difficult to detect them to meet the different needs coming from various demand sides. In this paper, we present an unsupervised solution for identifying rising stars in technological fields by mining patent information. The solution integrates three distinct aspects including technology performance, sociability and innovation caliber to present the profile of inventor, meantime, we design a series of features to reflect multifaceted ‘potential’ of an inventor. All features in the profile can get weights through the Entropy weight method, furthermore, these weights can ultimately act as the instruction for detecting different types of rising technology stars. A K-Means algorithm using clustering validity metrics automatically groups the inventors into clusters according to the strength of each inventor’s profile. In addition, using the nth percentile analysis of each cluster, this paper can infer which cluster with the most potential to become which type of rising technology stars. Through an empirical analysis, we demonstrate various types of rising technology stars: (1) tech-oriented RT Stars: growth of output and impact in recent years, especially in the recent 2 years; active productivity and impact over the last 5 years; (2) social-oriented RT Stars: own an extended co-inventor network and greater potential stemming from those collaborations; (3) innovation-oriented RT Stars: Various technical fields with strong innovation capabilities. (4) All-round RT Stars: show prominent potential in at least two aspects in terms of technical performance, sociability and innovation caliber.
LanguageEnglish
Pages137-172
Number of pages36
JournalScientometrics
Volume121
Issue number1
Early online date3 Aug 2019
DOIs
Publication statusPublished - 31 Oct 2019

Fingerprint

Stars
innovation
sociability
patent
Innovation
network analysis
entropy
performance
social network
productivity
instruction
Electric network analysis
demand
Clustering algorithms
resources
Entropy
Productivity
Group

Keywords

  • rising technology stars
  • co-inventor networks
  • social potential
  • technology performance
  • innovation caliber
  • tech mining

Cite this

Zhu, Lin ; Zhu, Donghua ; Wang, Xuefeng ; Cunningham, Scott W. ; Wang, Zhinan. / An integrated solution for detecting rising technology stars in co-inventor networks. In: Scientometrics. 2019 ; Vol. 121, No. 1. pp. 137-172.
@article{f9d7e62526504e538e637f85b59d28a6,
title = "An integrated solution for detecting rising technology stars in co-inventor networks",
abstract = "Online patent databases are powerful resources for tech mining and social network analysis and, especially, identifying rising technology stars in co-inventor networks. However, it’s difficult to detect them to meet the different needs coming from various demand sides. In this paper, we present an unsupervised solution for identifying rising stars in technological fields by mining patent information. The solution integrates three distinct aspects including technology performance, sociability and innovation caliber to present the profile of inventor, meantime, we design a series of features to reflect multifaceted ‘potential’ of an inventor. All features in the profile can get weights through the Entropy weight method, furthermore, these weights can ultimately act as the instruction for detecting different types of rising technology stars. A K-Means algorithm using clustering validity metrics automatically groups the inventors into clusters according to the strength of each inventor’s profile. In addition, using the nth percentile analysis of each cluster, this paper can infer which cluster with the most potential to become which type of rising technology stars. Through an empirical analysis, we demonstrate various types of rising technology stars: (1) tech-oriented RT Stars: growth of output and impact in recent years, especially in the recent 2 years; active productivity and impact over the last 5 years; (2) social-oriented RT Stars: own an extended co-inventor network and greater potential stemming from those collaborations; (3) innovation-oriented RT Stars: Various technical fields with strong innovation capabilities. (4) All-round RT Stars: show prominent potential in at least two aspects in terms of technical performance, sociability and innovation caliber.",
keywords = "rising technology stars, co-inventor networks, social potential, technology performance, innovation caliber, tech mining",
author = "Lin Zhu and Donghua Zhu and Xuefeng Wang and Cunningham, {Scott W.} and Zhinan Wang",
year = "2019",
month = "10",
day = "31",
doi = "10.1007/s11192-019-03194-w",
language = "English",
volume = "121",
pages = "137--172",
journal = "Scientometrics",
issn = "0138-9130",
number = "1",

}

An integrated solution for detecting rising technology stars in co-inventor networks. / Zhu, Lin; Zhu, Donghua; Wang, Xuefeng; Cunningham, Scott W.; Wang, Zhinan.

In: Scientometrics, Vol. 121, No. 1, 31.10.2019, p. 137-172.

Research output: Contribution to journalArticle

TY - JOUR

T1 - An integrated solution for detecting rising technology stars in co-inventor networks

AU - Zhu, Lin

AU - Zhu, Donghua

AU - Wang, Xuefeng

AU - Cunningham, Scott W.

AU - Wang, Zhinan

PY - 2019/10/31

Y1 - 2019/10/31

N2 - Online patent databases are powerful resources for tech mining and social network analysis and, especially, identifying rising technology stars in co-inventor networks. However, it’s difficult to detect them to meet the different needs coming from various demand sides. In this paper, we present an unsupervised solution for identifying rising stars in technological fields by mining patent information. The solution integrates three distinct aspects including technology performance, sociability and innovation caliber to present the profile of inventor, meantime, we design a series of features to reflect multifaceted ‘potential’ of an inventor. All features in the profile can get weights through the Entropy weight method, furthermore, these weights can ultimately act as the instruction for detecting different types of rising technology stars. A K-Means algorithm using clustering validity metrics automatically groups the inventors into clusters according to the strength of each inventor’s profile. In addition, using the nth percentile analysis of each cluster, this paper can infer which cluster with the most potential to become which type of rising technology stars. Through an empirical analysis, we demonstrate various types of rising technology stars: (1) tech-oriented RT Stars: growth of output and impact in recent years, especially in the recent 2 years; active productivity and impact over the last 5 years; (2) social-oriented RT Stars: own an extended co-inventor network and greater potential stemming from those collaborations; (3) innovation-oriented RT Stars: Various technical fields with strong innovation capabilities. (4) All-round RT Stars: show prominent potential in at least two aspects in terms of technical performance, sociability and innovation caliber.

AB - Online patent databases are powerful resources for tech mining and social network analysis and, especially, identifying rising technology stars in co-inventor networks. However, it’s difficult to detect them to meet the different needs coming from various demand sides. In this paper, we present an unsupervised solution for identifying rising stars in technological fields by mining patent information. The solution integrates three distinct aspects including technology performance, sociability and innovation caliber to present the profile of inventor, meantime, we design a series of features to reflect multifaceted ‘potential’ of an inventor. All features in the profile can get weights through the Entropy weight method, furthermore, these weights can ultimately act as the instruction for detecting different types of rising technology stars. A K-Means algorithm using clustering validity metrics automatically groups the inventors into clusters according to the strength of each inventor’s profile. In addition, using the nth percentile analysis of each cluster, this paper can infer which cluster with the most potential to become which type of rising technology stars. Through an empirical analysis, we demonstrate various types of rising technology stars: (1) tech-oriented RT Stars: growth of output and impact in recent years, especially in the recent 2 years; active productivity and impact over the last 5 years; (2) social-oriented RT Stars: own an extended co-inventor network and greater potential stemming from those collaborations; (3) innovation-oriented RT Stars: Various technical fields with strong innovation capabilities. (4) All-round RT Stars: show prominent potential in at least two aspects in terms of technical performance, sociability and innovation caliber.

KW - rising technology stars

KW - co-inventor networks

KW - social potential

KW - technology performance

KW - innovation caliber

KW - tech mining

U2 - 10.1007/s11192-019-03194-w

DO - 10.1007/s11192-019-03194-w

M3 - Article

VL - 121

SP - 137

EP - 172

JO - Scientometrics

T2 - Scientometrics

JF - Scientometrics

SN - 0138-9130

IS - 1

ER -