Vladimir Smojver, Mario Štorga and Goran Zovak
This paper aims to present a methodology by which future knowledge flow can be predicted by predicting co-citations of patents within a technology domain using a link prediction…
Abstract
Purpose
This paper aims to present a methodology by which future knowledge flow can be predicted by predicting co-citations of patents within a technology domain using a link prediction algorithm applied to a co-citation network.
Design/methodology/approach
Several methods and approaches are used: a dynamic analysis of a patent citation network to identify technology life cycle phases, patent co-citation network mapping from the patent citation network and the application of link prediction algorithms to the patent co-citation network.
Findings
The results of the presented study indicate that future knowledge flow within a technology domain can be predicted by predicting patent co-citations using the preferential attachment link prediction algorithm. Furthermore, they indicate that the patent – co-citations occurring between the end of the growth life cycle phase and the start of the maturation life cycle phase contribute the most to the precision of the knowledge flow prediction. Finally, it is demonstrated that most of the predicted knowledge flow occurs in a time period closely following the application of the link – prediction algorithm.
Practical implications
By having insight into future potential co-citations of patents, a firm can leverage its existing patent portfolio or asses the acquisition value of patents or the companies owning them.
Originality/value
It is demonstrated that the flow of knowledge in patent co-citation networks follows a rich get richer intuition. Moreover, it is show that the knowledge contained in younger patents has a greater chance of being cited again. Finally, it is demonstrated that these co-citations can be predicted in the short term when the preferential attachment algorithm is applied to a patent co-citation network.
Details
Keywords
Mario Štorga, Ali Mostashari and Tino Stanković
The paper aims to provide a methodology by which organisational knowledge can be extracted and visualised dynamically over time, providing a glimpse into the knowledge evolution…
Abstract
Purpose
The paper aims to provide a methodology by which organisational knowledge can be extracted and visualised dynamically over time, providing a glimpse into the knowledge evolution processes that occur within organisations.
Design/methodology/approach
Recursive analysis of email interactions is offered as a case to account for the knowledge structure evolution related to the different programs of international non-governmental organization (INGO). Several methods are used: analysis of the network expansion to see whether the process is random or uniform is performed, visualisation of the network configuration changes throughout studied time period; and the statistical examination of network formation.
Findings
The results of the presented study indicate that content structure of electronic knowledge networks exhibits hierarchical and centralised tendencies. The social network analysis results suggest that INGO exhibits non-hierarchical and decentralized structure of the individuals contributing to the discussion lists.
Research limitations/implications
By providing the means to carry out network evolution analysis of content structure dynamics and social interactions, the presented work provides a means for probabilistically modelling patterns of organisational knowledge evolution.
Practical implications
The approach allows the exploration of the dynamics of tacit to explicit knowledge, from individual to the group and from informal groups to the whole organisation.
Originality/value
By displaying the large collection of the key phrases that reflected the evolution of the organisational knowledge structure over the time, organisational emails are placed in meaningful context explaining the language of the organisation and context of knowledge structure evolution.