E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 118 Submitted 04/2018 Accepted for publication 11/2018 European Integration Studies No. 12 / 2018 pp. 118-128 DOI 10.5755/j01.eis.0.12.21869 Data-based Startup Profile Analysis in the European Smart Specialization Strat- egy: A Text Mining Approach EIS 12/2018 Data-based Startup Profile Analysis in the European Smart Specialization Strategy: A Text Mining Approach http://dx.doi.org/10.5755/j01.eis.0.12.21869 Levan Bzhalava Big Data Excellence Center, Kazimieras Simonavicius University Caucasus School of Business, Caucasus University Abstract Jari Kaivo-oja Finland Futures Research Centre, Turku School of Economics, University of Turku Sohaib S. Hassan SME Management Graduate School, University of Siegen The aim of the paper is to develop novel scientific metrics approach to the European Smart Specializa- tion Strategy. The European Union (EU) has introduced Smart Specialization Strategy (S3) to increase the innovation and competitive potential of its member states by identifying promising economic areas for investment and specialization. While the evaluation of Smart Specialization Strategy requires mea- surable criteria for the comparison of rate and level of development of countries and regions, policy makers lack efficient and viable tools for mapping promising sectors for smart specialization. To cope with this issue, we used a text mining approach to analyze the business description of startups from Nordic and Baltic countries in order to identify sectors in which entrepreneurs from these regions see new business opportunities. In particular, a topic modeling, Latent Dirichlet Allocation approach is employed to classify business descriptions and to identify sectors, in which start-up entrepreneurs identify possibilities of smart specialization. The results of the analysis show country-specific differ- ences in national startup profiles as well as variations among entrepreneurs coming from developed and less developed EU regions in terms of detecting business opportunities. Finally, we present policy implications for the European Smart Specialization Strategy. KEYWORDS: European Union, Smart Specialization Strategy (S3), S3 Implementation Handbook, Text Min- ing, Entrepreneurship, The Entrepreneurial Discovery Process (EDP) cycle, Innovation, opportunity search. 119 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 Introduction Smart Specialization Strategy (S3) is a new concept in territorial development (Foray et al., 2009; 2011; Boschma, 2016; Kaivo-oja et al., 2017). In particular, it is a place-based approach to stra- tegic economic growth and development, in which regions define their strengths and focus on finding niche areas of specialization in a global value chain (McCann and Ortega-Argilés, 2015). Smart Specialization concept places great importance on geographical context (social, cultural and institutional characteristics) in elaborating industrial and innovation policies (Foray et al., 2009; 2011; Barca at al., 2012). Given that regions differ in terms of their business and research activities, they compete in various technology and product spaces and show different strengths and weaknesses. Correspondingly, they have different opportunities for growth and development (McCann and Ortega-Argilés 2015). Moreover, knowledge spillovers are geographically bounded and they are rooted in place (Arrow, 1962; Audretsch and Feldman, 2004; Balland et al., 2018). In other words, technological knowledge that is complex and tacit in nature is difficult to imitate and is sticky in space. This, in turn, can be a primary source for competitive advantage for terri- tories in which they are generated (Balland et al., 2018). For these reasons, Smart Specialization Strategy (S3) shifts attention from traditional a ‘one-size-fits-all’ policy framework towards more embedded and locally relevant innovation policy. The smart specialization approach is considered as a key instrument to promote smart, sus- tainable and inclusive economic growth in Europe (Foray et al., 2011; Paliokaitė et al., 2016). Facing increased global competition, the European Union (EU) strives to encourage its regions to define their unique capabilities and to identify areas of specialization in which they can have competitive advantages. By enabling each region to identify a niche market of specialization, the EU aims to avoid duplication and fragmentation of its investments and to make its research and innovation efforts as effective as possible. In order to tap into an endogenous potential of each re- gion, Smart Specialization Strategy (S3) suggests that local actors from business and academia should discover the right areas of future specialization (Foray et al., 2009). In other words, it is a bottom-up approach in which local actors explore scientific and technological opportunities and their market potential. To build critical mass in the promising areas of specialization, regional governments should identify where local business and academia see potential for research and innovation activities and whether these activities show promise for excellence. This process, in turn, requires a deep analysis of local capabilities and competencies to identify unique features and strengths of each region and, based on this, to set priorities in innovation policies and to de- velop a regional vision. However, there is lack of clarity and consensus how to measure regional capabilities and competencies and how to implement Smart Specialization Strategy in practice. Previous studies suggest measuring Smart Specialization Strategy based on patent and indus- try analysis (Paliokaitė et al., 2016; Santoalha, 2016; Kaivo-oja et al., 2017; Asheim et al., 2017; Balland et al., 2018), but both these approaches have drawbacks in defining indices of smart specialization. First of all, not all inventions are patentable and, for this reason, patent analysis lacks indices to measure knowledge capabilities and competencies in a region. Industry analysis based on structured data collected by statistical offices may also be inappropriate for formulating smart specialization strategy, because smart specialization policy should target certain activities instead of industries and firms. Furthermore, innovations introduced by startup companies often disrupt value network in a given industry and displaces incumbent firms from a market as well as make previous technological knowledge obsolete (Christensen, 1997; Kaivo-oja and Laurae- us, 2018). In contrast to prior studies in smart specialization, we analyze the business descrip- tion of startup companies to identify activities in which entrepreneurs from different European regions see new business opportunities. Entrepreneurship is an important engine of economic growth and development (Wennekers and E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 120 Thurik, 1999; Zacharakis et al., 1999). Specifically, it is considered a source of innovative and competitive power of an economy (Wennekers and Thurik, 1999; Zacharakis et al., 1999; Praag and Versloot, 2007). As Wennekers and Thurik (1999) suggest, entrepreneurs play an indispen- sable role in economic growth because their activities create variety of ideas and initiatives. As a result, “variety, competition, selection and also imitation (...) expand and transform the pro- ductive potential of a regional or national economy” (Wennekers and Thurik, 1999: 50). This im- plies that entrepreneurs facilitate the operation of market selection mechanisms and promote innovation activities as well as stimulate industry evolution. For these reasons, entrepreneurial activities explain a substantial portion of variations in rates of economic growth across regions (Zacharakis et al.,1999). Entrepreneurial activities can be highly geographical and it may depend on local economic and institutional characteristics (Boschma, 2016). Moreover, entrepreneurial search process for busi- ness opportunities can be substantially shaped by local research and development activities (Acs et al., 2013). In other words, the asymmetries in knowledge accessibility across individuals can determine their capabilities to deliver high-valued solutions to market problems (Shane, 2003; Acs et al., 2013). As skills and know-how (e.g. tacit knowledge) are geographical bounded, en- trepreneurs across regions may develop different business opportunities and identify diverse areas of specialization. In this line of reasoning, we use a text mining approach to analyze the business description of startup companies from Nordic and Baltic countries in order to identify key economic areas in which entrepreneurs from these regions see new business opportunities. Specifically, a topic modeling approach is employed to classify full-text business descriptions and to identify economic activities, in which start-up entrepreneurs identify possibilities of smart specialization. The rest of the paper is organized in the following way. Section 2 reviews the related literature and provides theoretical framework. Section 3 presents the dataset and empirical methods used in the study. Section 4 discusses the findings from the empirical analysis and, at the end, section 5 summarizes and concludes. Smart specialization is a knowledge-driven growth strategy (Foray et al., 2009; 2011). In par- ticular, it is an innovation policy which aims to identify promising economic areas in a region for investment and specialization (Foray at al., 2011; Kaivo-oja et al., 2017). The European Union (EU) introduced Smart Specialization Strategy (S3) to help its member states discover the right sectors and fields of future Specializations in which they can have competitive advantages and, in this way, to increase their innovation and competitive potential (Foray et al., 2009; Foray et al., 2012; Kaivo-oja et al., 2017; Roman and Nyberg, 2017). Given that EU lacks economic and tech- nology specialization as well as has low capability to prioritize innovation and research efforts at regional level (Benner, 2013), adopting to Smart Specialization Strategy (S3) can support EU to develop innovative and resilient economy by specializing and concentrating on each region’s research and innovation strengths and, at the same time, avoiding duplication and fragmentation of investment efforts (Foray et al., 2012). Smart Specialization Strategy (S3) is a novel approach to regional economic development poli- cy (Kaivo-oja et al., 2017). The major idea that differentiates Smart Specialization Strategy (S3) from traditional industrial and innovation policy is its emphasis on a place-based and bottom-up approach in priority-setting (Foray at al., 2011; Boschma, 2016; Paliokaitė et al., 2016). In tradi- tional industrial and innovation policy, a decision making process in mapping promising sectors for investment and specialization was mainly centralized and top-down (Barca at al., 2012). In other words, decision makers were in a position to select and prioritize areas of economic ac- Theoretical framework 121 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 tivity with the highest growth potential and then to focus the development of clusters and inno- vation activities in these areas. Given that many uncertainties are involved in setting priorities in innovation policies and governments lack sufficient knowledge and expertise to design and implement such policies effectively, top-down sectoral approaches often fail to promote local economic development (Barca et al., 2012; McCann and Ortega-Argilés, 2016). In particular, “the evidence from numerous development policy examples worldwide demonstrates that regions have made many mistakes in terms of their policy choices, and often this was because policies were chosen on the basis of criteria which were not appropriate or relevant for the local context” (McCann and Ortega-Argilés, 2016: 282). To put it another way, top-down policy mostly assumes that a ‘one-size-fits-all’ policy framework is generally effective and relied on imitating successful innovation policies applied in very different contexts (Storper, 1997; Pike et al., 2006; Barca et al., 2012; McCann and Ortega-Argilés, 2016). Moreover, top-down centralized-organized policy fails to engage with small local actors and to reflect their interests in the policy design due to lack of their lobbying power, whereas major local players with dominant monopoly positions are able to influence the top-down policy formulation and to shape it in their own interests (Foray, 2015). In contrast to traditional industrial and innovation policy, Smart Specialization Strategy lets entre- preneurs and small local actors to discover the right areas of future specialization (Foray et al., 2011). Specifically, Smart Specialization Strategy relies Entrepreneurial Discovery Process (EDP) in setting innovation policy priorities (Foray at al., 2011; Coffano and Foray, 2014). Entrepreneuri- al Discovery Process is a bottom-up approach in which local actors from business and academia are discovering new market niches as well as scientific and technological opportunities (McCann and Ortega-Argilés, 2016; Boschma, 2016). The role of government in this process is to identify those entrepreneurial discovery projects or new activities and to develop clusters and innovation activities in these prioritized areas (Foray et al., 2011). Hence, the information necessary for set- ting priorities in a smart specialization perspective comes from local actors such as firms, lab- oratories and specialized services. This process intends “to allow innovation policies to emerge which are ‘placebased’, which build on a sound analysis of each region’s strengths and potential for excellence, and which involve a broad range of actors and their knowledge of market poten- tial” (Boschma, 2016: 17). Although a broad range of actors are involved in Entrepreneurial Dis- covery Process, entrepreneurs have a prominent role in this process (Coffano and Foray, 2014; Rodríguez-Pose and Wilkie, 2015), because they possess a valuable understanding of market dynamics and the commercial feasibility of scientific research activities due to their interaction with the market (Cities Alliance, 2007). Moreover, entrepreneurs serve as “agents of change” by facilitating the operation of market selection mechanisms and forcing established companies to become more productive and competitive as well as by promoting innovation activities and stim- ulating industry evolution. Specifically, their activities create variety of ideas and initiatives, and rivalry between these different ideas and initiatives lead to the selection of the most innovative and productive companies and the displacement of obsolete ones (Wennekers and Thurik, 1999). In other words, innovative entrepreneurs create new markets and industries by disrupting exist- ing structures, value and actors. As a result, entrepreneurs provide a significant contribution to the innovative and competitive power of a region by generating novel solutions to market prob- lems and renewing economic activities (Wennekers and Thurik, 1999; Zacharakis et al., 1999; Fölster, 2000; Praag and Versloot, 2007). The ability to identify complex problems and to deliver high-valued solutions is a key component of entrepreneurship (Shane, 2003). It allows entrepreneurs to recognize valuable business opportunities and, through a creative combination of resources, to bring into existence new products and services. Discovery of entrepreneurial opportunities can be considered as re- E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 122 combination activities, because an innovation is a new combination of the existing knowledge (Schumpeter, 1934). This implies that individuals operating in a knowledge-rich environment are more likely to develop business opportunities than others acting in a knowledge-impoverished context. The knowledge output in a given region is considered to be the function of R&D curried out by local universities and industries (Jaffe, 1986), and all the knowledge created by research institutions within a region is a potential source of entrepreneurial opportunities (Plummer and Acs, 2014). As knowledge spillovers are localized and knowledge spillovers are more intensive in regions with higher R&D investments (Audretsch and Feldman, 1996), regions vary in terms of their pool of knowledge and the knowledge-based opportunities generated for entrepreneurial exploitation (Acs et al., 2013). Therefore, identifying where entrepreneurs see business opportu- nities within a region can be an important instrument for defining local strengths and potential as well as developing a vision of regional smart specialization. To set priorities in a smart special- ization perspective, policy makers should focus not on individual entrepreneurial initiatives but a large set of them that can provide systematic processes in regional economic development. “Even though the entrepreneurial discovery is related to a micro level (individual initiatives that may result in new business projects), the approach of smart specialization model seeks to over- come it to reach a macro level” (Del Castillo Hermosa et al., 2015: 10). For this purpose, we ana- lyze business descriptions of individual early stage enterprises and aggregate them to provide a macro picture where entrepreneurs from Nordic and Baltic countries see business opportunities. Source: https://angel.co/europe Table 1 Number of startup companies in Nordic and Baltic Countries in 2016 and 2017 Data and methodology The empirical analysis of the paper is based on startup dataset of Nordic and Baltic countries. Specifically, we collected publicly available business descriptions of startup companies from the following website https://angel.co/europe, which provides information about startup activities from all over the world. In the empirical analysis, we focus on variations among entrepreneurs coming from developed and less developed EU regions in terms of detecting business opportu- nities. Global Entrepreneurship Monitor (GEM) study differentiates economic development levels among factor-driven, efficiency-driven and innovation-driven. Given that no factor-driven regions are presented in EU, we analyze business descriptions of early stage startup companies from efficiency-driven (Estonia, Latvia, Lithuania) and innovation-driven economies (Denmark, Fin- land, Sweden). According to GEM study, early stage entrepreneurs are those started business activities in the last 24 months. Therefore, we restricted our dataset to only companies started operations in 2016 and 20171 (Please see Table 1). Economic development level Country Number of Startups Efficiency-driven Estonia 227 Latvia 108 Lithuania 103 Innovation-driven Denmark 298 Finland 291 Sweden 283 For analyzing unstructured texts that describe businesses of startup companies, we use topic modeling technique. Topic modeling is an unsupervised machine learning method which au- tomatically clusters words that frequently occur together and discovers the abstract “topics”. In 1 Startup data for Denmark available only from June 2016 and for Sweden only from September 2016. 123 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 particular, given a collection of documents, topic modeling treats each document as a vector of word counts and present it as the mixture of topics, where each topic consists of relevant words. In the analysis, we use Latent Dirichlet Allocation of topic modeling algorithm (Blei et al., 2003; Shi et al., 2016), which associates business descriptions from each country to defined number of topics and each topic to a set of relevant keywords. Before applying the topic modeling algorithm, we used natural language processing methods to clean business description of startup compa- nies, which include removing stop words and punctuations as well as stemming words (reducing each word to a single root word). Estonia Rank Word Share 1 servic 38% 2 provid 33% 3 compani 32% 4 custom 30% 5 platform 29% Latvia Rank Word Share 1 servic 14% 2 design 14% 3 product 14% 4 app 12% 5 digit 11% Lithuania Rank Word Share 1 app 23% 2 develop 22% 3 design 21% 4 web 21% 5 product 20% Denmark Rank Word Share 1 platform 33% 2 product 28% 3 custom 27% 4 develop 26% 5 compani 23% Finland Rank Word Share 1 servic 23% 2 mobil 18% 3 game 16% 4 platform 16% 5 learn 14% Sweden Rank Word Share 1 platform 38% 2 app 33% 3 servic 32% 4 help 30% 5 develop 29% Table 2 The most popular keywords in startup company descritpions in Baltic and Nordic countries ResultsWe use natural language processing to identify which words are most often used in descrip-tions of startup companies in Nordic and Baltic economies. Table 2 shows that user experi- ence-centric words like service, app, design and platform are the most popular keywords in startup descriptions. In particular, early-stage entrepreneurs from Sweden (29%), Lithuania (23%) and Latvia (12%) refer to an app in their business descriptions. The term “service” is also mentioned frequently in descriptions for startups from Estonia (38%), Latvia (14%), Fin- land (23%) and Sweden (29%) as well as the word “platform” in descriptions of early-stage companies from Denmark (33%), Sweden (30%), Estonia (29%) and Finland (16%). Lithuania and Latvian companies in our dataset also frequently mention the term “design” in their de- scription (21% and 14%, respectively). In Finnish startup businesses, the words “mobile” and “game” are among the top five referred terms (see Table 2). E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 124 Estonia Latvia Lithuania Denmark Finland Sweden Figure 2 Word cloud of startup company business descritpions Looking at word cloud of startup company descriptions - 50 most popular keywords in startup company descriptions, Figure 1 shows that following terms app, platform, software, design, ser- vice, solution, data, market, digital, web, product and mobile appear often in startup landscape description of represented Nordic and Baltic countries. In qualitative analysis, these terms in- dicate key trends in start-up ecosystems in Nordic and Baltic countries. In general, the most popular keywords in startup company descriptions provide broad picture, but they do not show where startup companies from these economies see business opportunities. For this reason, we use topic modeling to classify business descriptions and to identify economic activities, in which start-up entrepreneurs identify business opportunities and possibilities of smart specialization. 125 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 Estonia Topic 1 Topic 2 Topic 3 Topic4 Topic 5 Topic 6 Topic 7 Topic 8 1 car solut service service team time provide market 2 custom provid truck data develop service user manag 3 compani busi cryptocurr provide app custom project provide 4 rental time blockchain product manag student service platform Latvia Topic 1 Topic 2 Topic 3 Topic4 Topic 5 Topic 6 Topic 7 Topic 8 1 service design reality real experi service market product 2 softwar digit world build fit financi manag hour 3 cryptocurr develop app industri smart app digit employe 4 custom first augment construct tool use agenc engag Lithuania Topic 1 Topic 2 Topic 3 Topic4 Topic 5 Topic 6 Topic 7 Topic 8 1 web start estat user train develop home product 2 design game real busi tool inform accessori improv 3 market employe service product product design app realiti 4 develop search claim deliveri swimmer hous artist online Denmark Topic 1 Topic 2 Topic 3 Topic4 Topic 5 Topic 6 Topic 7 Topic 8 1 busi develop content team service event product new 2 people platform art experi product app busi develop 3 custom compani design product custom compani app system 4 develop custom platform market time employe world manag Finland Topic 1 Topic 2 Topic 3 Topic4 Topic 5 Topic 6 Topic 7 Topic 8 1 service compani service digit mobil artifici inform product 2 smart softwar app busi game learn manag natur 3 home service car platform app educ communic server 4 busi develop pet will develop intellig mobil food Sweden Topic 1 Topic 2 Topic 3 Topic4 Topic 5 Topic 6 Topic 7 Topic 8 1 digit platform game servic data app product servic 2 music manag develop offer solut user system help 3 app use brand people smart help market platform 4 market develop experi need product publish platform compani Table 3 Topic modeling of startup company descritpions in Baltic and Nordic countries E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 126 Topics discovered through Latent Dirichlet Allocation of topic modeling algorithm are displayed in Table 3. Overall, Latent Dirichlet Allocation algorithm shows that there are some differences and similarities in the startup landscape in Nordic and Baltic countries. Looking first at efficien- cy-driven economies, early-stage Estonian entrepreneurs identify business opportunities in car rental service (Topic 1), cryptocurrency and blockchain related service (Topic 3), and data-based business solution service (Topic 2 and 4) as well as in developing project and team management (Topic 5 and 7), and market management apps (Topic 8). Somewhat similarly, start-up entre- preneurs from Latvia see new business opportunities in creating cryptocurrency and customer service software (Topic 1), digital design development (Topic 2), real estate and financial sectors (Topic 4 and 6), digital market management (Topic 7) and employee engagement service (Topic 8), and also in developing augmented reality apps (Topic 3). Lithuania companies are also active in real estate service, web design and development as well as in home design and accessories. As to the startup landscape of innovation-driven economies, companies from Denmark focus on developing event and employee management apps (Topic 6) as well as system development and management apps (Topic 8). Moreover, they are active in developing art and content design platforms (Topic 3) and also customer service and management platforms (Topic 1, 2, 5, 7). In Finland, early-stage entrepreneurs detect opportunities in smart home business (Topic 1), soft- ware service development (Topic 2 and 4), mobile game app development (Topic 5) as well as in developing apps for car rent, food service and also related to pets (Topic 3 and 8). Furthermore, in contrast to other countries, Finnish entrepreneurs identify possibilities of smart specialization in developing artificial intelligence for learning and education. Looking at Swedish startup com- panies, they are active in creating digital music apps (Topic 1), providing data smart business solutions (Topic 5), developing games (Topic 3) and market management platforms (Topic 7). Conclusions _ In this study, we explored the business descriptions of startup companies to get insight about startup landscape in Nordic and Baltic countries. We used Latent Dirichlet Allocation algo- rithm to discover economic activities in which entrepreneurs from these countries start new businesses and detect possibilities of smart specialization. The result suggests that there are some similarities and differences in startup communities in Nordic and Baltic countries. Our analysis shows that Smart Specialization Strategy approach and start-up community can be integrated by using text mining approaches. This methodological approach provides a lot of promising opportunities for European S3 research. For innovation ecosystem analyses text mining provides new understanding. _ This paper is first attempt to develop automated tools that can help governments to map prom- ising sectors for smart specialization without human intervention. Next steps are to measure competitiveness of startup activities across EU regions and to link entrepreneurial activities with patents, trademarks and brands generated in a region. Also focused data analyses to Industry 4.0 and Manufacturing 4.0 production structures would provide new insights to chang- ing innovation ecosystems in Europe. Understanding the Entrepreneurial Discovery Process (EDP) in the European Union would benefit much from large data-based text-mining analyses. References Acs, Z.J, Audretsch, D.B and Lehmann, E.E. (2013). The knowledge spillover theory of entrepreneur- ship. Small Business Economics, 41(4), 757-774. https://doi.org/10.1007/s11187-013-9505-9 Arrow, K. (1962). Economic Welfare and the Allo- cation of Resources for Invention. In R. Nelson (ed.) The Rate and Direction of Inventive Activity. Princ- eton University Press, Princeton, 609-626. https:// doi.org/10.1515/9781400879762-024 Asheim, B., Grillitsch, M. and Trippl, M. (2017). Smart specialization as an innovation-driven strat- egy for economic diversification: Examples from 127 E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 Scandinavian regions. In: Radosevic, S., Curaj, A., Gheorghiu, R., Andreescu, L. and Wade, I. (Eds.), Ad- vances in the Theory and Practice of Smart Special- ization. Academic press, Elsevier, 74-96. https:// doi.org/10.1016/B978-0-12-804137-6.00004-8 Audretsch, D. and Feldman, M. (1996). R&D Spill- overs and the Geography of Innovation and Pro- duction. The American Economic Review, 86(3): 630-640. Audretsch, D. and Feldman, M. (2004). Knowl- edge spillovers and the geography of innova- tion. Handbook of Regional and Urban Eco- nomics, 4, 2713-2739. https://doi.org/10.1016/ S1574-0080(04)80018-X Balland, P.A. Boschma, R., Crespo, J. and Rigby, D.L. (2018). Smart specialization policy in the Eu- ropean Union: relatedness, knowledge complexity and regional diversification. Regional Studies, DOI: 10.1080/00343404.2018.1437900. Barca, F., McCann, P. and Rodríguez-Pose, A. (2012). The case for regional development intervention: Place-based versus place-neutral approach- es. Journal of Regional Science, 52(1), 134-152. https://doi.org/10.1111/j.1467-9787.2011.00756.x Benner, M. (2013). From Smart Specialisation to Smart Experimentation: Towards a New Theo- retical Framework for EU Regional Policy. MPRA Paper No. 51843. https://mpra.ub.uni-muenchen. de/51843/1/MPRA_paper_51843.pdf Blei, D., Ng, A., and Jordan, M. (2003). Latent Dir- ichlet allocation. Journal of Machine Learning Re- search, 3, 993-1022. Boschma, R. (2016). Smart Specialisation and Re- gional Innovation Policy. Welsh Economic Review, 24, 17. DOI: http://doi.org/10.18573/j.2016.10050 Cities Alliance. (2007). Organising the Effort. In Understanding Your Local Economy - A Resource Guide for Cities (pp. 5–16). The Cities Alliance, Washington D.C. Coffano, M. and Foray, D. (2014). The Centrality of Entrepreneurial Discovery in Building and Imple- menting a Smart Specialisation Strategy. Scienze Regionali, 13(1), 33–50. https://doi.org/10.3280/ SCRE2014-001003 Christensen, C.M. (1997). The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail. Harvard Business School Press, Boston. Del Castillo Hermosa, J., Elorduy, J.P. and Eguía, B.B. (2015). Smart specialisation and entrepreneur- ial discovery: Theory and reality. The Portuguese Review of Regional Studies, 39, 6-22. Foray, D., David, P. A., and Hall, B. (2009). Smart Specialisation: The Concept. In Knowledge for Growth: Prospects for Science, Technology and In- novation. European Commission, Brussels. Foray, D., David, P.A., and Hall, B. (2011). Smart Specialisation: From Academic Idea to Political In- strument, the Surprising Career of a Concept and the Difficulties Involved in its Implementation. MTEI Working Paper No. 2011.001. https://pdfs.semantic- scholar.org/29ad/6773ef30f362d7d3937c483003d- 974bc91c5.pdf. Accessed April 2, 2018. Foray, D., Goddard, J., Beldarrain, X. G., Landaba- so, M., McCann, P., Morgan, K., Nauwelaers, C. and Ortega-Argiles, R. (2012). Guide to Research and Innovation Strategies for Smart Specialisations (RIS 3). Brussels: European Commission. http:// ec.europa.eu/regional_policy/sources/docgener/ presenta/smart_specialisation/smart_ris3_2012. pdf Accessed April 11, 2018. Foray, D. (2015). Smart Specialisation: Opportuni- ties and Challenges for Regional Innovation Policy. Routledge, London. Fölster, S. (2000). Do entrepreneurs create jobs? Small Business Economics, 14(2), 137–148. https://doi.org/10.1023/A:1008141516160 Jaffe, A. (1986). Technological Opportunity and Spillovers of R&D: Evidence from Firms’ Patents, Profits and Market Value. American Economic Re- view, LXXVI, 984-1001. Kaivo-oja, J. Vähäsantanen, S., Karppinen, A. and Haukioja, T. (2017). Smart specialization strategy and its operationalization in the regional policy: case Fin- land. Business, Management and Education, 15(1), 28–41. https://doi.org/10.3846/bme.2017.362 Kaivo-oja, J. and Lauraeus, T. (2018). The VUCA ap- proach as a solution concept to corporate foresight challenges and global technological disruption. Foresight, 20(1), 27-49. https://doi.org/10.1108/ FS-06-2017-0022 PLACE-BASED VERSUS PLACE-NEUTRAL AP- PROACHES McCann, P. and Ortega-Argilés, R. (2016). Smart specialisation: Insights from the EU experience and implications for other economies. Journal of Regional Research, 36, 279-293. McCann, P. and Ortega-Argilés, R. (2015). Smart Spe- cialisation, Regional Growth and Applications to EU Cohesion Policy. Regional Studies, 49(8), 1291-1302. https://doi.org/10.1080/00343404.2013.799769 Paliokaitė, A., Martinaitis, Z., and Sarpong, D. (2016). Implementing smart specialization road- maps in Lithuania: Lost in translation? Technolog- E u r o p e a n I n t e g r a t i o n S t u d i e s 2 0 1 8 / 1 2 128 ical Forecasting and Social Change, 110, 143-152. https://doi.org/10.1016/j.techfore.2016.01.005 Pike, A., Rodríguez-Pose, A. and Tomaney. J. (2006). Local and Regional Development. Routledge, Lon- don. https://doi.org/10.4324/9780203003060 Plummer, L.A. and Acs, Z.J. (2014). Localized competition in the knowledge spillover theory of entrepreneurship. Journal of Business Ventur- ing, 29(1), 121-136. https://doi.org/10.1016/j. jbusvent.2012.10.003 Praag, M. and Versloot, P.H. (2007). What is the value of entrepreneurship? A review of recent research. Small Business Economics, 29, 351-382.https://doi. org/10.1007/s11187-007-9074-x Roman, M. and Nyberg, T (2017). Smart Specialisa- tion Strategy Development in the Finnish Regions. Creating Conditions for Entrepreneurial Discovery. In K Panayiotis & S Adrian (eds), 10th International Conference for Entrepreneurship, Innovation and Re- gional Development ICEIRD 2017 – Conference Pro- ceedings, ICEIRD, pp. 363-370, International Confer- ence for Entrepreneurship, Innovation, and Regional Development, Thessaloniki, Greece, 31-1 September. Rodríguez-Pose, A. and Wilkie, C. (2015). Institu- tions and the Entrepreneurial Discovery Process for Smart Specialization. Papers in Evolutionary Economic Geography (PEEG) 1523, Utrecht Univer- sity, Department of Human Geography and Spatial Planning, Group Economic Geography, Utrecht, the Netherlands. http://econ.geo.uu.nl/peeg/ peeg1523.pdf. Accessed March 5, 2018. Santoalha, A. (2016). New Indicators of Smart Specialization: A related diversification approach applied to European Regions. Working Papers on Innovation Studies 20161220, Centre for Technolo- gy, Innovation and Culture, University of Oslo, Oslo. http://www.sv.uio.no/tik/InnoWP/tik_working paper_20161220.pdf. Accessed March 8, 2018. Schumpeter, J. A. (1934). The theory of economic development: An inquiry into profits, capital, credit, interest, and the business cycle, transl. by Redvers Opie, Harvard University Press, Harvard. Shane, S. (2003). A General Theory of Entrepre- neurship. Edward Elgar Publishing. Northamp- ton, MA: Edward Elgar Publishing. https://doi. org/10.4337/9781781007990 Shi, Z.M., Lee, G. and Whinston, A.B. (2016). Toward a Better Measure of Business Proximity: Topic Modeling for Industry Intelligence. MIS Quarterly, 40(4), 1035- 1056. https://doi.org/10.25300/MISQ/2016/40.4.11 Storper, M. (1997). The Regional World: Territorial Development in a Global Economy. Guilford Press, New York. Wennekers, S. and Thurik, R. (1999). Linking en- trepreneurship and economic growth. Small Business Economics, 13(1), 27–56. https://doi. org/10.1023/A:1008063200484 Zacharakis, A, Reynolds, P.D. and Bygrave, W.D. (1999). National Entrepreneurship Assessment: United States of America. Mo.: Kauffman Center for Entrepreneurial Leadership, Kansas City, MO. http://citeseerx.ist.psu.edu/viewdoc/download?- doi=10.1.1.194.8412&rep=rep1&type=pdf. Ac- cessed March 7, 2018. LEVAN BZHALAVA Ph.D. Big Data Excellence Center, Kaz- imieras Simonavicius University and Caucasus School of Business, Caucasus University Fields of research interests Innovation research, entrepreneur- ship, foresight research, business intelligence and Big Data analytics Address Dariaus ir Girėno g. 21, Vilnius 02189, Lithuania Tel. +370 633 93 0 47 Paata Saakadze 1, Tbilisi, 0102, Georgia Tel. +995 599 199 450 E-mails: levan.bzhalava@ksu.lt; lbzhalava@cu.edu.ge About the authors SOHAIB S. HASSAN Ph.D. SME Management Graduate School, University of Siegen, Germany Fields of research interests International Management, Innova- tion, SMEs, Regional Economics Address Unteres Schloss 3, 57072 Siegen, Germany Tel. +49-271-7402424 E-mail: sohaib.hassan@uni-siegen.de JARI KAIVO-OJA Ph.D., Prof. Finland Futures Research Cen- tre, Turku School of Econom- ics, University of Turku Fields of research interests Futures studies, foresight research, innovation research, Big Data and data analyt- ics, sustainable developme, globalisation, management, leadership, AI, Industry 4.0, economic growth, R&D poli- tics, start-up companies Address Åkerlundinkatu 2, 33100, Tampere, Finland Tel. +358-41-753 0244 E-mail: jari.kaivo-oja@utu.fi