download0 view151
twitter facebook

공공누리This item is licensed Korea Open Government License

Title
Identifying product opportunities using collaborative filtering based patent analysis
Author(s)
윤장혁이재민고병열서원철송인석
Publication Year
2016-04-19
Abstract
One practical and low-risk approach to product planning for technology-based firms is to identify applicationproducts based on their existing product portfolios. Previous studies, however, have tended toneglect the current product development capabilities of target firms and to apply the technical data ofspecific fields to their methods, thereby failing to quantify a way of identifying various product opportunities.As a remedy, this paper proposes a new multi-step approach to product recommendation. Thesteps include (1) generating assignee–product portfolio vectors using text mining on a large-scale sampleof patents, (2) recommending untapped products for a target firm by using latent Dirichlet allocation andcollaborative filtering, (3) producing a visual map based on the promise and domain heterogeneity of therecommended products. To validate the practicability, we applied our approach to a Korean high-techmanufacturer by using all of the patents registered in the United States Patent and Trademark Officedatabase during the period of time from 2009 to 2013. This study contributes to the systematic discoveryof new product opportunities across various domains using the existing product portfolios of firms, andcould become the basis for a future product opportunity analysis system.
Keyword
Product opportunity; Product portfolio; Collaborative filtering; Latent Dirichlet allocation; Text mining; Patent analysis
Journal Title
Computers & Industrial Engineering
ISSN
0360-8352
Files in This Item:
There are no files associated with this item.
Appears in Collections:
7. KISTI 연구성과 > 학술지 발표논문
URI
https://repository.kisti.re.kr/handle/10580/14535
Export
RIS (EndNote)
XLS (Excel)
XML

Browse