شناخت فن آوری محوری مبتنی بر اثرات متقابل تکنولوژیکی: استخراج قانون رابطه (ARM) و رویکرد فرایند شبکه تحلیلی (ANP)

This study proposes a new approach to identifying core technologies from a perspective of technological cross-impacts based on patent co-classification information with consideration of the overall interrelationships among technologies. The proposed approach is comprised of two methods: association rule mining (ARM) and the analytic network process (ANP). Firstly association rule mining (ARM) is employed to calculate the technological cross-impact indexes. Since the confidence measure in ARM is defined as a conditional probability between two technologies, it is adopted as an index for evaluating technological cross-impacts. The technological cross-impact matrix is then constructed with all calculated cross-impact indexes. Secondly, the ANP, which is a generalization of the analytic hierarchy process (AHP), is conducted to produce priorities of technologies with consideration of their direct and indirect impacts. The proposed approach can be utilized for technology monitoring for both technology planning of firms and innovation policy making of governments. A case of telecommunication technology is presented to illustrate the proposed approach.

The characteristics of modern technological changes can be defined as complexity and radicalness. Under such environment, it has become more important to grasp technological trends and advances by analyzing the overall structure of technologies and interaction among them (Lee, Kim, Cho, & Park, 2009). It is considered to be an indispensible activity, in particular, for seeking technological possibilities through technological fusion among various fields of technologies such as IT, BT, and NT. Consequently, there have often been attempts to identify technological structure and relationship. The identification of technological structure and relationship is mainly conducted through the patent analysis (Trajtenberg, 1990). It is reported that patents contain about 80% of all technological knowledge (Blackman, 1995), and they can be easily accessed and analyzed through various types of public or private databases. Patents are, hence, perceived as useful information for techno-economic analysis and R&D management (Yoon & Park, 2004). A lot of studies have attempted to analyze technological relationships using patent information. The most commonly used information of patents for analyzing technological relationships is citation information. The basic assumptions of citation analysis are that the knowledge of cited patent is transferred to a citing patent, and there exists a technological linkage between them. Citation analysis is a useful method for identifying technological relationships, and this can be verified with various studies (Basberg, 1987, Hu and Jaffe, 2003, Jaffe and Trajtenberg, 1999, Trajtenberg, 1990 and Yoon and Park, 2004). However, they have some shortcomings reported in the literature. First, the average time-lag between citing and cited patents is over 10 years (Hall, Jaffe, & Trajtenberg, 2001). Moreover, since citation analysis only considers citing-cited relationships between individual patents, it is difficult to identify technological relatedness and characteristics from a perspective of technological fields (Yoon & Park, 2004). To address this limitation, there have been attempts to applying other pieces of information of patents such as co-citation (Lai and Wu, 2005 and Stuart and Podoly, 1996), co-word (Courtial, Callon, & Sigogneau, 1993), and keyword vector (Yoon & Park, 2004) for the analysis of technological relationship. They also have, however, their own weaknesses. There is still a time-lag problem in the co-citation analysis. Co-word analysis and keyword vector analysis require qualitative judgments, and therefore have a lack of consistency in the result of analysis. On the contrary, the patent analysis with co-classification information has some advantages over the above mentioned measures. Co-classification analysis is to analyze technological relationships based on the fact that patents are classified to some technological classes considering their technological characteristics (OECD, 1994). The assumption of the co-classification analysis is that the frequency by which two classification codes are jointly assigned to a patent document can be interpreted as a sign of the strength of the knowledge relationships, in terms of knowledge links and spillovers (Breschi, Lissoni, & Maleraba, 2003). In contrast to the citation analysis, it is based on the hierarchical technological classification system so that technological relationships can be analyzed not on the level of individual patents but on the various technological levels according to the purpose of studies. Furthermore, errors from the time-lag problem are relatively insignificant since the time of classification information of a patent is equal – patent registration time. Among the various techniques using the information of patent co-classification, technological cross-impact analysis (CIA) has been used to identify core technologies and interrelationships between technologies by analyzing the cross-impact between technologies quantitatively based on patent classification data (Choi, Kim, & Park, 2007). In the patent-based CIA, the cross-impact index between two technologies is calculated with the probabilities based on the patent co-classification information to analyze the impact between technologies. However, it is subject to some limitations. Firstly, it is nearly impossible to calculate cross-impact indexes without developing a computer program because it requires a huge amount of calculation with patent data. Secondly, regarding to the identification of core technologies, the previous patent-based CIA does not take into account the overall interrelationships among technologies, only considers the relationships between two technologies. In response, this paper proposes a new approach to identifying core technologies from a perspective of cross-impacts based on patent co-classification information with consideration of the overall interrelationships among technologies. The proposed approach is comprised of two methods: association rule mining (ARM) and the analytic network process (ANP). At first, association rule mining (ARM) is employed to calculate technological cross-impact indexes. ARM is one of the representative data mining techniques for exploring vast database. Since the confidence measure in ARM is defined as a conditional probability between two technologies and is of the same formula with cross-impact index, it is adopted as the index of evaluating technological cross-impacts. Then, the technological cross-impact matrix is constructed with all calculated cross-impact indexes. Secondly, the ANP, which is a generalization of the analytic hierarchy process (AHP), is employed to identify core technologies based on the cross-impact matrix. Since the ANP is capable of measuring the relative importance that captures all the indirect interactions in a network, the derived limit priority indicates the importance of a technology in terms of impacts on other technologies, taking all the direct and indirect influences into account (Lee et al., 2009). The remainder of the paper is organized as follows: Section 2 deals with methodological background including CIA, ARM, and the ANP. The proposed approach is explained in Section 3, and illustrated in Section 4. The paper ends with conclusions in Section 5.

This study has proposed a systemic approach to identification of core technologies from a perspective of the technological cross-impacts. ARM was applied to the patent co-classification data, and the technological cross-impact matrix was constructed with the derived confidence value of each technology. Then, the ANP was employed to prioritize technologies from a perspective of the overall interrelationship among technologies. For the purpose of illustration, a case example of telecommunication technology was presented. This research contributes to the field by proposing a new approach to identification of core technologies based on the overall technological cross-impacts. ARM is employed to more easily capture the technological cross-impacts among technologies, and the ANP is then conducted to produce priorities of technologies with consideration of their direct and indirect impacts. The proposed approach can be utilized for technology monitoring for both technology planning of firms and innovation policy making of governments. Nonetheless, this study is still subject to some limitations, and these limitations will serve as fruitful areas of further research. Firstly, since the proposed approach was illustrated at the class level, the ANP is restricted to only clusters with no elements. Extending the analysis to the subclass level of patents could make use of the full potential of the ANP. An extension of analysis to all technologies in ICT could also be considered as future research issues. Finally, the selected 13 patent classes as telecommunication technologies are by no means exhaustive. A more systematic procedure is required to select the target classes.