THE PROBLEM

Due to the nanotechnology’s substantial impact on numerous economic sectors, more than 60 countries have investedin national projects or programs to stimulate research and innovation in nanotechnology (Dang, Chen, Zhang, & Roco, 2009).In 2000, the National Nanotechnology Initiative (NNI) was established by the U.S. government to encourage investment in nanotechnology (Huang, Chen, Li, & Roco, 2006). The cumulative NNI investment now amounts to around $14 billion since its inception, including the 2011 budget of $1.8 billion(Funding, 2011), and the investment is on the increase every year.

In order to make the NNI investment effective, it is necessary to identify what kinds of patents orresearch topics will be dominant in the near future.According to Huang’s research (Huang, Chen, Li, & Roco, 2006), the topics of nanotechnology changed significantlyafter the year of 2000,and it is also expected that thisradical change of topics will go on, meaning that the NNI investment mustreflect this change in topics. Therefore, the purposes of this research are to show that public funding from NNI has significantly contributed to the development of nanotechnology, and to predict what research topics will be dominant in the near future.

LITERATURE REVIEW

kNOWLEDGE DIFFUSION IN NANOTECHNOLOGY

Patents have been used to assess the evolution and development of topics in nanotechnology(Dang, Chen, Zhang, & Roco, 2009), so patents are considered to represent knowledge. Knowledge diffusion can be defined as the adaptations and applications of knowledge which is documented in scientific publications and patents (Chen & Hicks, 2004).In other words, the emergence and development of scientific knowledge is attributable to a diffusion process in which new ideas are spreadamong researchers(Chen & Hicks, 2004).Citing means adopting knowledge elements from cited scientific publications or patents, and human collaboration leads to knowledge evolution and innovation by combining prior knowledge elements(Li, Chen, Zhang, Li, & Jay, 2009), so citations of research papers representthe collaboration among researchers.

EPIDEMIC THEORY FOR KNOWLEDGE DIFFUSION

In 1964, Goffman adopted the Epidemic theory (Kermack & McKendrick, 1925) to explain the transmission of idea based on the assumption that ideas are transmitted by personal communication or by publication (Goffman & Newill, 1964). One advantage of the epidemic theory is the ability to estimate population (Bettencourt, Clintron-Arias, Kaiser, & Castillo-Chavez, 2006), so that it leads to more rational analysis of why and how infectious diseases begin and continue and what factors lead to the decline and disappearance of diseases.

THE SIR MODEL

People are affected or unaffected to other peoples’ ideas or knowledge. When person is affected to ideas of other people, he/she may try to talk about the ideas with other people, resulting in a knowledge ‘epidemic’ (Goffman & Newill, 1964). The SIR model, which represents Susceptible, Infected, and Recovered, is devised to explain how an infectious disease spreads(Kermack & McKendrick, 1925). Goffman (Goffman & Newill, 1964)described an epidemic process as “The ‘epidemic’ process can be characterized as one of transition from on state (susceptible) to another (infective) where the transition is caused by exposure to some phenomenon (infectious material).”

  • S: susceptibles - those who can become infectives given contact with infectious material.
  • I: infected - those members who are host to the infectious material.
  • R: Recovered - those who have been removed for one for a variety of reasons such as death, immunity, hospitalization etc.

Figure 1: SIR Model(Goffman & Newill, 1964)

The table below clearly shows the analogy between infectious disease epidemic and knowledge diffusion in nanotechnology.

Comparison
Compartments
of SIR Model / Epidemics / Knowledge Diffusion
(Information Diffusion)
About / Disease / Knowledge / Topic (of Nanotechnology)
Susceptible / People who might make a contact with an infective / Inventors who might conduct research on specific topics of nanotechnology after reading a certain patent information
Infected / People who have disease and might infect others / Inventors who conduct research on specific topics of nanotechnology and make citations on their research works
Recovered / People who are recovered from disease / Inventors who no longer conduct research on specific topics of nanotechnology
Infection Rate / The probability of disease transmission from the susceptible to the infected / The probability of conducting research on specific topics of on nanotechnology after reading a certain patent information related with one’s research interests
Recovery Rate / The probability that the infected is recovered from disease / The probability that inventors who conducted research on specific topics of nanotechnology lose their research interests

Table 1: Analogy between disease epidemic and knowledge diffusion

RESEARCH QUESTIONS

From the prior studies, we were able to obtain clear evidence that the evolution and development of research topics in nanotechnology can be explained by the epidemic model, that is, the SIR model.Then, our focus is placed how well the model explainsthe lifecycle of topics.

The proposed SEIR-F model will be compared with the baseline SIR model in terms of predicting accuracy. They are all designed to predict the population dynamics, and by using the dynamics, the epidemic model allows us to estimate whichtopics will emerge, live, and die. Especially, the SEIR-F model takes into account public funding, so it might be expectedthat the SEIR-F predicts better than the SIR model. Research questions are followed:

  • Can the epidemic theory (the SIR / SEIR-F models) explain knowledge diffusion in nanotechnology?
  • Have public funding facilitated research activities in nanotechnology?
  • Can the epidemic theorypredict the evolution, development, and disappearance of specific topics?

METHODOLOGYAND RESEARCH APPROACH

EXPERIMENT STEPS

Figure 2: Experiment Steps

  • Data acquisition: The USPTO patent dataset is already acquired. The dataset covers data from 1976 through 2009.
  • Preparation: In order to construct citation network, graph-based algorithms will be used, so necessary data will be extracted and prepared for constructing citation network.
  • Technique: To identify topics in patent content, keywords in patent content are extracted by Arizona Noun Phraser and then they will be clustered for identifying topics by Self Organizing Map (SOM).
  • Analysis: The SIR and the SEIR-F models will be used to explain knowledge diffusion in nanotechnology, and predict the evolution and development of specific topics.

TOPIC AND SOM ANALYSES

Topic analysis will be conducted for identifying the major research topics in nanotechnology. For example, Arizona Noun Phraser (Li, Chen, Dang, & Lin, 2008)designed for extracting noun phrases will be performedfor identifyingresearch topics represented by a set of keywords.Then Self Organizing Map (SOM) will produce clusters based on extracted keywords. After getting clusters, we will label each cluster based on keywords.

THE SEIR-F MODEL

c(Bettencourt, Clintron-Arias, Kaiser, & Castillo-Chavez, 2006).

Figure 3: SIR Model

The second reason is that public funding facilitates idea adoption (research progress) by providing necessary resources such as specific equipment, laboratory materials, human resources, or so.As a result, the SEIR-F model is designed to take into account two reasons mentioned above.

RESULTS (Expected)

Knowledge diffusion in nanotechnology is well described by the SIR and the SEIR-F models. As I expected, the SEIR-F model outperformed the SIR model in terms of describing numbers of researchers who adopt, transmit, and develop new ideas, meaning that public funding has significantly contributed to the nanotechnology fields.The predictability of emerging and developing topicswill allow NNI to invest onright topics in a timely fashion.

Bibliography

Funding. (2011). Retrieved October 9, 2011, from National Nanotechnology Initiative:

Bettencourt, L. M., Clintron-Arias, A., Kaiser, D. I., & Castillo-Chavez, C. (2006). The power of a good idea: Quantitative modeling of the spread of ideas from epidemiological models. Physica A, 513-536.

Chen, C., & Hicks, D. (2004). Tracing knowledge diffusion. Scientometrics, 199-211.

Dang, Y., Chen, H., Zhang, Y., & Roco, M. C. (2009, 9). Knowledge sharing and diffusion patterns. pp. 9-21.

Goffman, W., & Newill, V. A. (1964). Generalization of Epidemic Theory. Nature, 225-204.

Huang, Z., Chen, H., Li, X., & Roco, M. C. (2006). Connecting NSF Funding to Patent Innovation in Nanotechnology (2001-2004). Journal of Nanoparticle Research, 859-879.

Kermack, W., & McKendrick, A. (1925). Applications of mathematics to medical problems. Edinburgh Mathematical Society, 1-34.

Li, X., Chen, H., Dang, Y., & Lin, Y. (2008). A longitudinal analysis of nanotechnology literature: 1976-2004. Journla of Nanoparticle Research, 3-22.

Li, X., Chen, H., Zhang, Z., Li, J., & Jay, N. F. (2009). Managing knowledge in light of its evolution process: An empirical study on citation netwokr-based patent classification. Journal of Management Information Systems, 129-153.