Patent data analysis using functional count data model

Jong Min Kim, University of Minnesota Morris
Nak Kyeong Kim, Virginia Commonwealth University
Yoonsung Jung, Prairie View A and M University
Sunghae Jun, Cheongju University


Technology is an important cause of social change. So many researchers have studied on diverse methods for technology analysis. Patent analysis has been proposed in many studies for technical analysis. They extracted technological keywords and codes from patent documents and analyzed them using statistics and machine learning. One of the problems in the existing studies was the patent analysis that did not consider the time factor. However, time is a factor to be considered in technology analysis. Because technology has evolved over time, in this paper, we study and propose a new technology analysis method considering time factor. We analyze patent data to understand technological structure of company, because patent contains most of information about developed technology. A lot of studies on technology analysis using patent data have been published in various areas. Many of them used extracted technological keywords from patent documents for patent analysis. They did not consider time factor to build technology analysis models, but we know technology changes over time. So we propose a technology analysis method using functional data analysis as a patent analysis considering time factor. We select Apple technology for our case study. With the patent data of Apple over time, we investigate on the technological structure of Apple and its technological evolution through high-dimensional visualization using harmonic components generated by functional data analysis. In addition, by employing the count data regression models of Poisson, negative binomial and hurdle Poisson, we examine the relationships among highest frequency keywords based on the visual outputs in functional data analysis. The practical implication of this paper is that it can be applied more effectively than the existing studies of technology analysis by considering time factor. This research contributes to technology forecasting for understanding the social changing. We can develop a more efficient research and development plan to improve the technological competition. The originality of this research is to consider time factor in technology analysis based on patent data. In this paper, we used the functional data analysis to model trends of technology keywords over time. Using the results of the technology analysis of this study considering the time, the company will be able to understand the social change and thereby improve its technological competitiveness in the market.