三维打印技术的潜在风险识别:跨学科主题词挖掘视角
|
摘要
在智能制造和工业互联网的体系架构中,三维打印(增材制造)是重要的核心技术之一。但是,当前大部分针对三维打印的相关研究更多聚焦于创新和市场潜力。针对三维打印技术普及和产业化可能带来的健康、环境和生态风险等问题,相关研究则较为零散,而且分布在化学、材料、物理、生物和环境等不同专业和研究领域。这给相关的科技政策制定和预防式公共治理措施出台带来了一定的信息阻塞和决策障碍。此外,目前有关新兴技术潜在风险识别的理论和实证研究均较少,相关的量化分析尤甚,可以认为是当前科技评价和技术管理研究领域的盲点之一。针对当前有关新兴技术潜在风险识别与评估的相关科学问题,本文提出了一个基于朴素机器学习思想的跨学科主题词挖掘框架,并结合当前较为热点的三维打印技术进行了实证分析。理论和实证分析显示,本文提出的跨学科主题词挖掘框架对于某个具体新兴技术潜在风险的分析,具有一定的内容识别效果,这对于相关的公共治理决策和科技政策制定,或许具有一定的参考价值和积极意义。
In terms of the contemporary intelligent manufacturing( IM) and industrial Internet,three-dimensional printing( 3DP)/additive manufacturing( AM) is definitely one of the most important key technologies because 3DP/AM can bring the disruptive innovation from high-level design to low-level manufacturing. Therefore,the academic topics about 3DP/AM attract more and more attention from many different kinds of scholars and institutes. While we retrospect the relevant publications on 3DP/AM during the past decades,we can find that most of the current research on 3DP/AM focuses on the huge innovation and market potential of 3DP/AM,and very few studies ever mentioned or touched the possible or potential risk of 3DP/AM. Actually,the main streams of studies on 3DP/AM just inform the very positive impact on economic systems,and very little is known about the possible/potential risk or threat to health,environment and ecology,security and social ethic and so on. Once 3DP/AM technologies would be popularized and industrialized,besides the tremendous pleasure for the emerging opportunities of economic development,the corresponding public-policies are also very important for the social-economic systems; otherwise,the emerging technologies could cause a new negative case of technology commercialization.However,the related studies and publications to the risk of 3DP/AM are still very few,and this exploration onthe risk of 3DP/AM are scattered in many different categories or academic areas,for example,chemistry,materials science,physics,biology and environment science and so forth. Obviously,because the risk signals on 3DP/AM are so weak and difficult to be identified and captured by those non-proficient policy-makers,a certain degree of information blockage and decision-making obstacles unintentionally derived,and the formulation of relevant policies on anticipatory governance has to face the difficulty of knowledge-acquisition. Further,the theoretical explorations and empirical analyses on the potential risk of emerging technologies are rare in the past decades,in particular the relevant quantitative studies on risk of a specific emerging technology almost keep blank; even could be a blind spot in the relevant areas on emerging technologies.To mitigate the gap between contemporary study on risk of emerging technology and concrete requirements of anticipatory governance and public policy making,a quantitative analytical method to identify the risk signal of a specific emerging technology is proposed,which can be attributed to a mining algorithm of cross-topic terms,and depends on the philosophy of nave machine learning. In the proposed framework of identifying the risk signals from the relevant publications on the observed emerging technology,the quantitative method can be divided into such different steps as:( a) the preparation of simple machine learning;( b) algorithm training based on a small sample;( c) the optimization of the relevant parameters. In each step,some concepts and computing equations are subsequently proposed.To verify the proposed identifying method for risk signals of emerging technology,the case study of 3DP/AM is conducted in empirical analysis. Relatively,the proposed method for identifying the risk of a specific emerging technology can provide moderate accuracy and recall ration,and significantly improve the efficiency of search the risk studies on a specific emerging to some degree. In terms of the empirical case on 3DP/AM,some important information is revealed and visualized that several significant risk or threat to human health and social security should be paid more attention in the contemporary stage of 3DP/AM development. First,several commercial polymer materials for 3DP in market present the biological toxicity that could make harmful impact on human health. Second,the ultrafine particles produced by 3DP/AM technologies not only bring the negative emission to the environment,but also bring the detrimental effect on human's lung,in particular in the office space,and the harmful effect of ultrafine mental or polymer particles could cause unpredictable diseases. Third,the highly concentrated usage of power/electricity in 3DP/AM technologies could bring the huge impact on the stability of grid because the sintering steps in 3DP/AM rely on laser technology. In addition,3DP/AM could aggravate the potential infringement of intellectual property because the technology of 3DP can make much easier copycats,and then could bring a big blow to those creative corporations and industries.Meanwhile,very few studies or researchers in social science can timely knew or responded to those risk signals of 3DP/AM,which could mean that the public concerns and public perceptions on 3DP/AM could be incomplete,i. e. the most of us just knew the huge potential for economic development about 3DP/AM,and very lack of the knowledge about the possible threats from 3DP/AM. In consequence,the relevant public-policies are highly possible time-lag,and even invalid while the health or security risk of 3DP/AM would become true events. China has the huge population,and very high density of population in some big cities; therefore,the potential or possible risk/threat of 3DP/AM or the other emerging technologies to human health,security and ethic should cause more attention from Chinese scholars,institutes and policy-makers because the loss in China or India could be much larger than the other small countries while those risk or threats of emerging technologies had really happened.However,the theoretical exploration and the empirical case on 3DP/AM also have limitations,for example,the mining method of trans-disciplinary topic terms has the room to enhance the accuracy,and whether the case study of 3DP/AM can be utilized by the other emerging technologies and so on. In addition,this paper just provides a new perspective to retrospect the prior studies on 3DP/AM,and does not imply the possible direction for the relevant researchers,i. e. how to encourage more scholars or institutes join the explorations on identifying risk signals for a specific emerging technology is not revealed by far.In view of the current scientific problems related to the identification and assessment of potential risks of emerging technologies,this paper proposes an interdisciplinary/trans-disciplinary topic word mining algorithm based on naive machine learning,and makes an empirical analysis with the current hot three-dimensional printing technology. Theoretical and empirical analysis show that the proposed method has a certain content recognition capability on the mining the potential risk signal for a specific emerging technology,which may have a certain value and positive significance for public governance decision-making and science and technology policy formulation for emerging technology.
In terms of the contemporary intelligent manufacturing( IM) and industrial Internet,three-dimensional printing( 3DP)/additive manufacturing( AM) is definitely one of the most important key technologies because 3DP/AM can bring the disruptive innovation from high-level design to low-level manufacturing. Therefore,the academic topics about 3DP/AM attract more and more attention from many different kinds of scholars and institutes. While we retrospect the relevant publications on 3DP/AM during the past decades,we can find that most of the current research on 3DP/AM focuses on the huge innovation and market potential of 3DP/AM,and very few studies ever mentioned or touched the possible or potential risk of 3DP/AM. Actually,the main streams of studies on 3DP/AM just inform the very positive impact on economic systems,and very little is known about the possible/potential risk or threat to health,environment and ecology,security and social ethic and so on. Once 3DP/AM technologies would be popularized and industrialized,besides the tremendous pleasure for the emerging opportunities of economic development,the corresponding public-policies are also very important for the social-economic systems; otherwise,the emerging technologies could cause a new negative case of technology commercialization.However,the related studies and publications to the risk of 3DP/AM are still very few,and this exploration onthe risk of 3DP/AM are scattered in many different categories or academic areas,for example,chemistry,materials science,physics,biology and environment science and so forth. Obviously,because the risk signals on 3DP/AM are so weak and difficult to be identified and captured by those non-proficient policy-makers,a certain degree of information blockage and decision-making obstacles unintentionally derived,and the formulation of relevant policies on anticipatory governance has to face the difficulty of knowledge-acquisition. Further,the theoretical explorations and empirical analyses on the potential risk of emerging technologies are rare in the past decades,in particular the relevant quantitative studies on risk of a specific emerging technology almost keep blank; even could be a blind spot in the relevant areas on emerging technologies.To mitigate the gap between contemporary study on risk of emerging technology and concrete requirements of anticipatory governance and public policy making,a quantitative analytical method to identify the risk signal of a specific emerging technology is proposed,which can be attributed to a mining algorithm of cross-topic terms,and depends on the philosophy of nave machine learning. In the proposed framework of identifying the risk signals from the relevant publications on the observed emerging technology,the quantitative method can be divided into such different steps as:( a) the preparation of simple machine learning;( b) algorithm training based on a small sample;( c) the optimization of the relevant parameters. In each step,some concepts and computing equations are subsequently proposed.To verify the proposed identifying method for risk signals of emerging technology,the case study of 3DP/AM is conducted in empirical analysis. Relatively,the proposed method for identifying the risk of a specific emerging technology can provide moderate accuracy and recall ration,and significantly improve the efficiency of search the risk studies on a specific emerging to some degree. In terms of the empirical case on 3DP/AM,some important information is revealed and visualized that several significant risk or threat to human health and social security should be paid more attention in the contemporary stage of 3DP/AM development. First,several commercial polymer materials for 3DP in market present the biological toxicity that could make harmful impact on human health. Second,the ultrafine particles produced by 3DP/AM technologies not only bring the negative emission to the environment,but also bring the detrimental effect on human's lung,in particular in the office space,and the harmful effect of ultrafine mental or polymer particles could cause unpredictable diseases. Third,the highly concentrated usage of power/electricity in 3DP/AM technologies could bring the huge impact on the stability of grid because the sintering steps in 3DP/AM rely on laser technology. In addition,3DP/AM could aggravate the potential infringement of intellectual property because the technology of 3DP can make much easier copycats,and then could bring a big blow to those creative corporations and industries.Meanwhile,very few studies or researchers in social science can timely knew or responded to those risk signals of 3DP/AM,which could mean that the public concerns and public perceptions on 3DP/AM could be incomplete,i. e. the most of us just knew the huge potential for economic development about 3DP/AM,and very lack of the knowledge about the possible threats from 3DP/AM. In consequence,the relevant public-policies are highly possible time-lag,and even invalid while the health or security risk of 3DP/AM would become true events. China has the huge population,and very high density of population in some big cities; therefore,the potential or possible risk/threat of 3DP/AM or the other emerging technologies to human health,security and ethic should cause more attention from Chinese scholars,institutes and policy-makers because the loss in China or India could be much larger than the other small countries while those risk or threats of emerging technologies had really happened.However,the theoretical exploration and the empirical case on 3DP/AM also have limitations,for example,the mining method of trans-disciplinary topic terms has the room to enhance the accuracy,and whether the case study of 3DP/AM can be utilized by the other emerging technologies and so on. In addition,this paper just provides a new perspective to retrospect the prior studies on 3DP/AM,and does not imply the possible direction for the relevant researchers,i. e. how to encourage more scholars or institutes join the explorations on identifying risk signals for a specific emerging technology is not revealed by far.In view of the current scientific problems related to the identification and assessment of potential risks of emerging technologies,this paper proposes an interdisciplinary/trans-disciplinary topic word mining algorithm based on naive machine learning,and makes an empirical analysis with the current hot three-dimensional printing technology. Theoretical and empirical analysis show that the proposed method has a certain content recognition capability on the mining the potential risk signal for a specific emerging technology,which may have a certain value and positive significance for public governance decision-making and science and technology policy formulation for emerging technology.
引文
[1]Rotolo D,Hicks D,Martin B R.What is an emerging technology?[J].Research Policy,2015,44(10):1827-1843.
[2]李牧南.新兴技术管理的热点研究分析-国内外主要相关文献的主题词突引分析视角[J].情报杂志,2016,35(11):45-50.Li Munan.Exploration on hot issues of emerging technology management:A burst-citation analysis of terms based on the main relevant literature[J].Journal of Intelligence,2016,35(11):45-50.
[3]高盼,邢冬梅.乌尔里希·贝克技术风险思想探析[J].科技管理研究,2017(12):260-266.Gao Pan,Xing Dongmei.The research on the thought of technology risk of Ulrich Beck[J].Science and Technology Management Research,2017(12):260-266.
[4]欧庭高,巩红新.论现代技术风险冲突及其综合治理[J].科技管理研究,2015(15):254-258.Ou Tinggao,Gong Hongxin.Modern technology risk conflict and management[J].Science and Technology Management Research,2015(15):254-258.
[5]Roca J B,Vaishnav P,Morgan M G,et al.When risks cannot be seen:Regulating uncertainty in emerging technologies[J].Research Policy,2017,46(7):1215-1233.
[6]Li C,Ramakrishna S,Singh S.A review of digital manufacturing-based hybrid additive manufacturing processes[J].International Journal of Advanced Manufacturing Technology,2018,95(5-8):2281-2300.
[7]Upadhyay M,Sivarupan T,Mansori M E.3D printing for rapid sand casting-A review[J].Journal of Manufacturing Processes,2017,29:211-220.
[8]Weller C,Kleer R,Piller F T.Economic implications of 3Dprinting:Market structure models in light of additive manufacturing revisited[J].International Journal of Production Economics,2015,164:43-56.
[9]Zhu F.,Friedrich T.,Nugegoda D.,et al.Assessment of the biocompatibility of three-dimensional-printed polymers using multispecies toxicity tests[J].Biomicrofluidics,2015,061103;doi:10.1063/1.4939031.
[10]WaltherG.Printing insecurity?The security implications of3D-Printing of weapons[J].Science and Engineering Ethics,2015,21(5):1435-1445.
[11]杰里米·里夫金(作者),张体伟(译者),孙豫宁(译者).第三次工业革命[M].北京:中信出版社,2012(6).Jeremy Rifkin(Author).Zhang Tiwei(Translator),Sun Yuning(Translator).The third industrial revolution[M].Beijing:Zhongxin Press,2012(6).
[12]乌尔里希·森德勒(Ulrich Sendler)(编者),邓敏(译者),李现民(译者).工业4.0:即将来袭的第四次工业革命[M].北京:机械工业出版社,2014(7).Ulrich Sendler(Author),Deng Min(Translator),Li Xianmin(Translator).Industry 4.0[M].Beijing:Machinery Industry Press,2014(7).
[13]Thwala M,Klaine S J,Musee N.Interactions of metalBased engineered nanoparticles with aquatic higher plants:Areview of the state of current knowledge[J].Environmental Toxicology and Chemistry,2016,35(7):1677-1694.
[14]Jeannet N,Fierz M,Schneider S,et al.Acute toxicity of silver and carbon nanoaerosols to normal and cystic fibrosis human bronchial epithelial cells[J].Nanotoxicology,2016,10(3):279-291.
[15]Youtie J,Porter A,Shapira P,et al.The use of environmental,health and safety research in nanotechnology research[J].Journal of Nanoscience and Nanotechnology,2011,11(1):158-166.
[16]Carley S,Porter A L.Measuring the influence of nanotechnology environmental,health and safety research[J].Research Evaluation,2011,20(5):389-395.
[17]Walker W C,Bosso C J,Eckelman M,et al.Integrating life cycle assessment into managing potential EHS risks of engineered nanomaterials:reviewing progress to date[J].Journal of Nanoparticle Research,2015,17(8):344.
[18]Zhou X,Zhang Y,Porter A L,et al.A patent analysis method to trace technology evolutionary pathways[J].Scientometrics,2014,100(3):705-721.
[19]Zhang Y,Zhou X,Porter A L,et al.How to combine term clumping and technology roadmapping for newly emerging science&technology competitive intelligence:"problem&solution"pattern based semantic TRIZ tool and case study[J].Scientometrics,2014,101(2):1375-1389.
[20]Zhang M Y,Li J,Hu H,et al.Seizing the strategic opportunities of emerging technologies by building up innovation system:Monoclonal antibody development in China[J].Health Research Policy and Systems,2015,13(1):01-12.
[21]Dong K,Xu H,Rui L,et al.An integrated method for interdisciplinary topic identification and prediction:A case study on information science and library science[J].Scientometrics,2018,115(2):849-868.
[22]Yau C K,Porter A,Newman N,et al.Clustering scientific documents with topic modeling[J].Scientometrics,2014,100(3):767-786.
[23]吴泽福,Jie WU.技术创新风险整合监管机制研究[J].科研管理,2017,38(6):18-27.Wu Zefu,Jie WU.A research on the integration and supervison mechanism of technology innovation risks[J].Science Research Management,2017,38(6):18-27.
[24]Blei,DM.Probabilistic topic models[J].Communications of the ACM,2012,55(4),77-84.
[25]Li M,Porter A L.Facilitating the discovery of relevant studies on risk analysis for three-dimensional printing based on an integrated framework[J].Scientometrics,2018,114(1):277-300.
[26]Qin P,Xu W,Guo J.A novel negative sampling based on TFIDF for learning word representation[J].Neurocomputing,2016,177:257-265.
[27]Li M.Classifying and ranking topic terms based on a novel approach:Role differentiation of author keywords[J].Scientometrics,2018,116(1):77-100.
[28]Diabat A,Batta6a O,Nazzal D.An improved lagrangian relaxation-based heuristic for a joint location-inventory problem[J].Computers&Operations Research,2015,61:170-178.
[29]Ghahramani Z.Probabilistic machine learning and artificial intelligence[J].Nature,2015,521(7553):452.
[30]朱军,胡文波.贝叶斯机器学习前沿进展综述[J].计算机研究与发展,2015,52(1):16-26.Zhu Jun,Hu Wenbo.Recent advances in bayesian machine learning[J].Journal of Computer Research and Development,2015,52(1):16-26.
[31]Li,MN.,&Chu,YQ.Explore the research front of a specific research theme based on a novel technique of enhanced co-word analysis[J].Journal of Information Science,2017,43(6):725-741.
[32]潘东华,徐珂珂.基于共词分析的技术机会分析[J].科研管理,2014,35(4):10-17.Pan Donghua,Xu Keke.Technology opportunity analysis based on co-word analysis[J].Science Research Management,2014,35(4):10-17.
[33]Brandes,Ulrik.A faster algorithm for betweenness centrality[J].Journal of Mathematical Sociology,2001,25(2):163-177.
[34]李牧南.技术预见研究热点的演进分析:内容挖掘视角[J].科研管理,2018,39(3):141-153.Li Munan.An analysis of the evolution of hot issues on technology foresight based on content mining[J].Science Research Management,2018,39(3):141-153.
[35]陈悦,陈超美,刘则渊,等.CiteSpace知识图谱的方法论功能[J].科学学研究,2015,33(2):242-253.Chen Yue,Chen Chaomei,Liu Zeyuan,et al.The methodology function of CiteSpace mapping knowledge domains[J].Studies in Science of Science,2015,33(2):242-253.
[36]Rong K,Patton D,Chen W.Business models dynamics and business ecosystems in the emerging 3D printing industry[J].Technological Forecasting&Social Change,2018,134(9):234-245.
[37]Melchels F P W,Feijen J,Grijpma D W.A review on stereolithography and its applications in biomedical engineering[J].Biomaterials,2010,31(24):6121-6130.
[38]Huang Y,Zhu D,Qian Y,et al.A hybrid method to trace technology evolution pathways:A case study of 3D printing[J].Scientometrics,2017,111(1):185-204.
[39]Afshar-Mohajer N,Wu C Y,Ladun T,et al.Characterization of particulate matters and total VOC emissions from a binder jetting 3D printer[J].Building and Environment,2015,93:293-301.
[40]Yi J,Le Bouf R F,Duling M G,et al.Emission of particulatematter from a desktop three-dimensional(3D)printer[J].Journal of Toxicology and Environmental Health,Part A,2016,79(11):453-465.
[41]Peng T.Analysis of energy utilization in 3d printing processes[J].Procedia CIRP,2016,40:62-67.
[42]Neely E L.The risks of revolution:Ethical dilemmas in 3Dprinting from a US perspective[J].Science&Engineering Ethics,2016,22(5):1285-1297.
[43]Chan F L,House R,Kudla I,et al.Health survey of employees regularly using 3D printers[J].Occupational Medicine,2018,68(3):211-214.
[44]De M.,Henning H.,Ferreira P.,et al.Potential health and environmental risks of three-dimensional engineered polymers[J].Environmental Science&Technology Letters,2018,5(2):80-85.
(1)反向文档频率样本数据的抽样方式如下:首先,按作者字母顺序排列(产生一定的随机分布效应);其次,用随机函数生成4个跨度为5000-10000的随机数,并以这四个随机数为起始编码获取500条纪录,则每年获取2000篇文献,2000-2016为17年,共计获取34000条文献(Article)。鉴于Web of Science每次只能下载最大500条纪录,这里的抽样依然不是完全随机方式,但是作为反向文档频率计算的参考样本,而样本数量达到34000,基本可以满足代表性要求。
(1)这里反向文档频率-IDF的权重较大主要是因为文献检索方式是基于“新兴技术风险分析”主题,因此采集的文献中与风险相关的主题词自然较多,因此频次分布呈现较大差异,为了平衡这种共现频次本身的差异,因此对基于随机样本的反向文档频率-IDF的赋权加大,之所以采取0. 2和0. 8的权重主要是基于通常意义上的“二八原理”,可能存在一定的争议。实际上,在本案例中普通TF*IDF加权与词频排序基本一致,也就是说词频之间的差距过大,导致普通(传统)反向文档频率(IDF)加权完全失效。
[2]李牧南.新兴技术管理的热点研究分析-国内外主要相关文献的主题词突引分析视角[J].情报杂志,2016,35(11):45-50.Li Munan.Exploration on hot issues of emerging technology management:A burst-citation analysis of terms based on the main relevant literature[J].Journal of Intelligence,2016,35(11):45-50.
[3]高盼,邢冬梅.乌尔里希·贝克技术风险思想探析[J].科技管理研究,2017(12):260-266.Gao Pan,Xing Dongmei.The research on the thought of technology risk of Ulrich Beck[J].Science and Technology Management Research,2017(12):260-266.
[4]欧庭高,巩红新.论现代技术风险冲突及其综合治理[J].科技管理研究,2015(15):254-258.Ou Tinggao,Gong Hongxin.Modern technology risk conflict and management[J].Science and Technology Management Research,2015(15):254-258.
[5]Roca J B,Vaishnav P,Morgan M G,et al.When risks cannot be seen:Regulating uncertainty in emerging technologies[J].Research Policy,2017,46(7):1215-1233.
[6]Li C,Ramakrishna S,Singh S.A review of digital manufacturing-based hybrid additive manufacturing processes[J].International Journal of Advanced Manufacturing Technology,2018,95(5-8):2281-2300.
[7]Upadhyay M,Sivarupan T,Mansori M E.3D printing for rapid sand casting-A review[J].Journal of Manufacturing Processes,2017,29:211-220.
[8]Weller C,Kleer R,Piller F T.Economic implications of 3Dprinting:Market structure models in light of additive manufacturing revisited[J].International Journal of Production Economics,2015,164:43-56.
[9]Zhu F.,Friedrich T.,Nugegoda D.,et al.Assessment of the biocompatibility of three-dimensional-printed polymers using multispecies toxicity tests[J].Biomicrofluidics,2015,061103;doi:10.1063/1.4939031.
[10]WaltherG.Printing insecurity?The security implications of3D-Printing of weapons[J].Science and Engineering Ethics,2015,21(5):1435-1445.
[11]杰里米·里夫金(作者),张体伟(译者),孙豫宁(译者).第三次工业革命[M].北京:中信出版社,2012(6).Jeremy Rifkin(Author).Zhang Tiwei(Translator),Sun Yuning(Translator).The third industrial revolution[M].Beijing:Zhongxin Press,2012(6).
[12]乌尔里希·森德勒(Ulrich Sendler)(编者),邓敏(译者),李现民(译者).工业4.0:即将来袭的第四次工业革命[M].北京:机械工业出版社,2014(7).Ulrich Sendler(Author),Deng Min(Translator),Li Xianmin(Translator).Industry 4.0[M].Beijing:Machinery Industry Press,2014(7).
[13]Thwala M,Klaine S J,Musee N.Interactions of metalBased engineered nanoparticles with aquatic higher plants:Areview of the state of current knowledge[J].Environmental Toxicology and Chemistry,2016,35(7):1677-1694.
[14]Jeannet N,Fierz M,Schneider S,et al.Acute toxicity of silver and carbon nanoaerosols to normal and cystic fibrosis human bronchial epithelial cells[J].Nanotoxicology,2016,10(3):279-291.
[15]Youtie J,Porter A,Shapira P,et al.The use of environmental,health and safety research in nanotechnology research[J].Journal of Nanoscience and Nanotechnology,2011,11(1):158-166.
[16]Carley S,Porter A L.Measuring the influence of nanotechnology environmental,health and safety research[J].Research Evaluation,2011,20(5):389-395.
[17]Walker W C,Bosso C J,Eckelman M,et al.Integrating life cycle assessment into managing potential EHS risks of engineered nanomaterials:reviewing progress to date[J].Journal of Nanoparticle Research,2015,17(8):344.
[18]Zhou X,Zhang Y,Porter A L,et al.A patent analysis method to trace technology evolutionary pathways[J].Scientometrics,2014,100(3):705-721.
[19]Zhang Y,Zhou X,Porter A L,et al.How to combine term clumping and technology roadmapping for newly emerging science&technology competitive intelligence:"problem&solution"pattern based semantic TRIZ tool and case study[J].Scientometrics,2014,101(2):1375-1389.
[20]Zhang M Y,Li J,Hu H,et al.Seizing the strategic opportunities of emerging technologies by building up innovation system:Monoclonal antibody development in China[J].Health Research Policy and Systems,2015,13(1):01-12.
[21]Dong K,Xu H,Rui L,et al.An integrated method for interdisciplinary topic identification and prediction:A case study on information science and library science[J].Scientometrics,2018,115(2):849-868.
[22]Yau C K,Porter A,Newman N,et al.Clustering scientific documents with topic modeling[J].Scientometrics,2014,100(3):767-786.
[23]吴泽福,Jie WU.技术创新风险整合监管机制研究[J].科研管理,2017,38(6):18-27.Wu Zefu,Jie WU.A research on the integration and supervison mechanism of technology innovation risks[J].Science Research Management,2017,38(6):18-27.
[24]Blei,DM.Probabilistic topic models[J].Communications of the ACM,2012,55(4),77-84.
[25]Li M,Porter A L.Facilitating the discovery of relevant studies on risk analysis for three-dimensional printing based on an integrated framework[J].Scientometrics,2018,114(1):277-300.
[26]Qin P,Xu W,Guo J.A novel negative sampling based on TFIDF for learning word representation[J].Neurocomputing,2016,177:257-265.
[27]Li M.Classifying and ranking topic terms based on a novel approach:Role differentiation of author keywords[J].Scientometrics,2018,116(1):77-100.
[28]Diabat A,Batta6a O,Nazzal D.An improved lagrangian relaxation-based heuristic for a joint location-inventory problem[J].Computers&Operations Research,2015,61:170-178.
[29]Ghahramani Z.Probabilistic machine learning and artificial intelligence[J].Nature,2015,521(7553):452.
[30]朱军,胡文波.贝叶斯机器学习前沿进展综述[J].计算机研究与发展,2015,52(1):16-26.Zhu Jun,Hu Wenbo.Recent advances in bayesian machine learning[J].Journal of Computer Research and Development,2015,52(1):16-26.
[31]Li,MN.,&Chu,YQ.Explore the research front of a specific research theme based on a novel technique of enhanced co-word analysis[J].Journal of Information Science,2017,43(6):725-741.
[32]潘东华,徐珂珂.基于共词分析的技术机会分析[J].科研管理,2014,35(4):10-17.Pan Donghua,Xu Keke.Technology opportunity analysis based on co-word analysis[J].Science Research Management,2014,35(4):10-17.
[33]Brandes,Ulrik.A faster algorithm for betweenness centrality[J].Journal of Mathematical Sociology,2001,25(2):163-177.
[34]李牧南.技术预见研究热点的演进分析:内容挖掘视角[J].科研管理,2018,39(3):141-153.Li Munan.An analysis of the evolution of hot issues on technology foresight based on content mining[J].Science Research Management,2018,39(3):141-153.
[35]陈悦,陈超美,刘则渊,等.CiteSpace知识图谱的方法论功能[J].科学学研究,2015,33(2):242-253.Chen Yue,Chen Chaomei,Liu Zeyuan,et al.The methodology function of CiteSpace mapping knowledge domains[J].Studies in Science of Science,2015,33(2):242-253.
[36]Rong K,Patton D,Chen W.Business models dynamics and business ecosystems in the emerging 3D printing industry[J].Technological Forecasting&Social Change,2018,134(9):234-245.
[37]Melchels F P W,Feijen J,Grijpma D W.A review on stereolithography and its applications in biomedical engineering[J].Biomaterials,2010,31(24):6121-6130.
[38]Huang Y,Zhu D,Qian Y,et al.A hybrid method to trace technology evolution pathways:A case study of 3D printing[J].Scientometrics,2017,111(1):185-204.
[39]Afshar-Mohajer N,Wu C Y,Ladun T,et al.Characterization of particulate matters and total VOC emissions from a binder jetting 3D printer[J].Building and Environment,2015,93:293-301.
[40]Yi J,Le Bouf R F,Duling M G,et al.Emission of particulatematter from a desktop three-dimensional(3D)printer[J].Journal of Toxicology and Environmental Health,Part A,2016,79(11):453-465.
[41]Peng T.Analysis of energy utilization in 3d printing processes[J].Procedia CIRP,2016,40:62-67.
[42]Neely E L.The risks of revolution:Ethical dilemmas in 3Dprinting from a US perspective[J].Science&Engineering Ethics,2016,22(5):1285-1297.
[43]Chan F L,House R,Kudla I,et al.Health survey of employees regularly using 3D printers[J].Occupational Medicine,2018,68(3):211-214.
[44]De M.,Henning H.,Ferreira P.,et al.Potential health and environmental risks of three-dimensional engineered polymers[J].Environmental Science&Technology Letters,2018,5(2):80-85.
(1)反向文档频率样本数据的抽样方式如下:首先,按作者字母顺序排列(产生一定的随机分布效应);其次,用随机函数生成4个跨度为5000-10000的随机数,并以这四个随机数为起始编码获取500条纪录,则每年获取2000篇文献,2000-2016为17年,共计获取34000条文献(Article)。鉴于Web of Science每次只能下载最大500条纪录,这里的抽样依然不是完全随机方式,但是作为反向文档频率计算的参考样本,而样本数量达到34000,基本可以满足代表性要求。
(1)这里反向文档频率-IDF的权重较大主要是因为文献检索方式是基于“新兴技术风险分析”主题,因此采集的文献中与风险相关的主题词自然较多,因此频次分布呈现较大差异,为了平衡这种共现频次本身的差异,因此对基于随机样本的反向文档频率-IDF的赋权加大,之所以采取0. 2和0. 8的权重主要是基于通常意义上的“二八原理”,可能存在一定的争议。实际上,在本案例中普通TF*IDF加权与词频排序基本一致,也就是说词频之间的差距过大,导致普通(传统)反向文档频率(IDF)加权完全失效。