基于模糊语义模型的陶瓷智能评估系统的研究
|
摘要
陶瓷在日常生活中无处不在,人们在购买时不仅要求价格适中,更要求质量可靠,使其具有较高的性价比。而目前,我国陶瓷行业仍然采用手感和目测等方法进行质量评估,其评估结果往往是不可靠的。因此建立一套完整的、可靠的陶瓷质量评估系统已成为一项迫切的工作和研究内容。感观评估是通过人的视觉、嗅觉、味觉、触觉、听觉所获信号的评估技术,目前已广泛应用于食品、化妆品、纺织、化学、包装、汽车等行业中,并已成为研究的新热点。在这种情况下,我们将感观评估技术引入到陶瓷行业中,建立起一套完整的陶瓷智能感观评估平台。
本文结合陶瓷行业自身特点,运用灰色理论、模糊语义、D-S证据理论、多元统计分析、遗传算法等理论和方法,以MATLAB和SPSS等应用软件为工具,建立了一个适用于陶瓷行业的感观评估系统。该系统主要由感观评估、客观评估、感观客观之间联系三部分组成。
在感观评估部分,首先,我们建立了一个数据标准化模型,并且通过实例验证了我们所建模型的有效性;其次,通过设计调查问卷表对普通消费者的评估数据进行收集,用统计学中的因子分析方法对收集来的数据进行筛选,选用K-均值聚类方法对筛选后的数据进行聚类分析;由于专家和普通消费者之间对评估词汇的定义和理解有所差异,很容易造成专家与普通消费者评估之间的冲突,因此,最后采用遗传算法建立模型,解决了专家评估与普通消费者评估之间的冲突问题。客观评估部分,结合核函数和主成分分析法对仪器测量的客观数据进行核主成分压缩,并通过灰色聚类方法对压缩后的数据进行聚类分析。此陶瓷感观评估体系已用于景德镇陶瓷企业,效果明显。
We can see domestic ceramic anywhere in daily life, when buying them people not only ask the advisable price but also ask the high quality, make it have good cost performance. But nowadays in our country handle and visual measurement are still adopted as the method of ceramics quality evaluation, the result sometimes is unreliable. So it is necessary to establish a dependable quality evaluation system for ceramic. Sensory evaluation is an evaluation which process signal attained by seeing、smell、taste、feeling、hearing. Now, it is widely used in the field of food industry、cosmetic industry、textile industry、chemical industry、packaging industry、automobile industry and so on. In such a case the technology of sensory evaluation was imported to the ceramic industry, we have established an intelligent sensory evaluation system for ceramic.
This paper established an intelligent sensory evaluation system for ceramic with the software MATLAB and SPSS based on Grey System Theory、Fuzzy Semantics、D-S evident theory、Multivariate Statistical Analysis and Genetic Algorithm. The system is made up of sensory evaluation, objective evaluation and the contact between sensory evaluation and objective evaluation.
In the part of sensory evaluation, first of all, a ceramic model of sensory data standardization was established and the validity was validated by a case; then, we collect evaluation date from ordinary customer by design principles of consumer questionnaire; next, we built a data filtration process by factorial analysis; also we propose a fuzzy c-mean clustering method; due to experts and ordinary consumers' difference to evaluation vocabulary' define and understanding which is likely to result in conflict between evaluation experts and ordinary consumers, finally we established a model to resolve it by genetic algorithm. In the part of objective evaluation, by combining the kernel function and principal component with it, we reduce the date which is measured by apparatus, while processing date with clustering analysis we propose a gray clustering method. The sensory evaluation is applied to ceramic companies in jingdezhen and the effect is distinct.
本文结合陶瓷行业自身特点,运用灰色理论、模糊语义、D-S证据理论、多元统计分析、遗传算法等理论和方法,以MATLAB和SPSS等应用软件为工具,建立了一个适用于陶瓷行业的感观评估系统。该系统主要由感观评估、客观评估、感观客观之间联系三部分组成。
在感观评估部分,首先,我们建立了一个数据标准化模型,并且通过实例验证了我们所建模型的有效性;其次,通过设计调查问卷表对普通消费者的评估数据进行收集,用统计学中的因子分析方法对收集来的数据进行筛选,选用K-均值聚类方法对筛选后的数据进行聚类分析;由于专家和普通消费者之间对评估词汇的定义和理解有所差异,很容易造成专家与普通消费者评估之间的冲突,因此,最后采用遗传算法建立模型,解决了专家评估与普通消费者评估之间的冲突问题。客观评估部分,结合核函数和主成分分析法对仪器测量的客观数据进行核主成分压缩,并通过灰色聚类方法对压缩后的数据进行聚类分析。此陶瓷感观评估体系已用于景德镇陶瓷企业,效果明显。
We can see domestic ceramic anywhere in daily life, when buying them people not only ask the advisable price but also ask the high quality, make it have good cost performance. But nowadays in our country handle and visual measurement are still adopted as the method of ceramics quality evaluation, the result sometimes is unreliable. So it is necessary to establish a dependable quality evaluation system for ceramic. Sensory evaluation is an evaluation which process signal attained by seeing、smell、taste、feeling、hearing. Now, it is widely used in the field of food industry、cosmetic industry、textile industry、chemical industry、packaging industry、automobile industry and so on. In such a case the technology of sensory evaluation was imported to the ceramic industry, we have established an intelligent sensory evaluation system for ceramic.
This paper established an intelligent sensory evaluation system for ceramic with the software MATLAB and SPSS based on Grey System Theory、Fuzzy Semantics、D-S evident theory、Multivariate Statistical Analysis and Genetic Algorithm. The system is made up of sensory evaluation, objective evaluation and the contact between sensory evaluation and objective evaluation.
In the part of sensory evaluation, first of all, a ceramic model of sensory data standardization was established and the validity was validated by a case; then, we collect evaluation date from ordinary customer by design principles of consumer questionnaire; next, we built a data filtration process by factorial analysis; also we propose a fuzzy c-mean clustering method; due to experts and ordinary consumers' difference to evaluation vocabulary' define and understanding which is likely to result in conflict between evaluation experts and ordinary consumers, finally we established a model to resolve it by genetic algorithm. In the part of objective evaluation, by combining the kernel function and principal component with it, we reduce the date which is measured by apparatus, while processing date with clustering analysis we propose a gray clustering method. The sensory evaluation is applied to ceramic companies in jingdezhen and the effect is distinct.
引文
[1]秦寿康等.综合评价原理与应用[M].北京:电子工业出版社,2003
[2]Dijksterhuis G B.Multivariate Data Analysis in Sensory and Consumer Science[M].Food & Nutrition Press Inc,Trumbull,Connecticut USA,1997.
[3]Ruan Da,The Role of Computational Intelligence in Sensory Evaluation[J].Journal of Donghua University(Eng,Dd.),2004.21(3):21-25.
[4]刘思峰,党耀国,方志耕等.灰色系统理论及其应用[M].北京:科学出版社,2004
[5]Ruan Da,ZENG Xian-y.iIntelligent Sensory Evaluation.Belgium[J].ENSAIT Textile Institute.2003:1-10.
[6]袁红春,钱权,丁德贵.神经网络技术在茶叶感官品质评定中的应用[J].中国茶叶,1999(1):14-15.
[7]周跃斌,尚本清,邓克尼.名优绿茶质量评定方法的研究[J].湖南农业大学学报,1997,23(2):145-149.
[8]赵玉香.感官分析机理与茶叶感官审评条件[J].中国茶叶加工,2001,4:41-42.
[9]覃金钱.感官分析在造纸行业中的应用[J].标准计量与质量,1997(6):37-38.
[10]王光华.感官分析在肉制品加工中的应用[J].肉类工业,1994(12):32-36.
[11]张爱霞,陆淳等.感官分析技术在食品工业中的应用[J].中国乳品工业,2005,33(3):39-40.
[12]张爱霞,邓宏斌等.感官分析技术及在食品工业中的应用[J].乳业科学与技术,2004(3):113-114.
[13]雍克岚,吕敬慈.烟草三维荧光特征和品质评吸的关系[J].化学研究与应用,1998,10(5):494-497.
[14]肖绚等.论陶艺创作中的符号学[J].中国陶瓷,2005(204):32-35.
[15]胡鸿豪,陈刚等.陶瓷智能感官评估平台的研究[J].中国陶瓷,2006(3):36-38
[16]白雪梅等.对主成分分析综合评价方法若干问题的探讨[J].统计学-经济数学方法,1996(1):47-51
[17]Pan N,Zeronian S H,Ryu H-S.An alternative approach to the objective of fabrics[J].Textile Research Journal,1993,63(1):33-43.
[18]Pan N,Yen K C,Zhao S J,et al.A new approach to the objectivemeasurement of fabrics handle from mechanics properties-Part Ⅰ[J].Textile Research Journal,1988,58(8):438-443.
[19]周林峰.基于智能感官评估的服装供应链信息流反馈系统的研究[硕士论文].上海:东华大学,2005
[20]GB/T3532-1995.日用瓷器[S].北京:中国标准出版社,1995
[21]Azeem MF,Hanmandlu M.Generaliantion of adaptive Neuro-fuzzy inference systems[J].IEEE Transactions On Neural Networks,2000
[22]李宏,徐晖,孙仲康.一种具有信息融合功能的模糊推理方法[J].系统工程与电子技术,1998(1):52-54.
[23]罗志增,蒋静坪.基于D-S理论的多信息融合方法及应用[J].电子学报,1999,27(9):100-102.
[24]张文修,梁怡.不确定性推理原理[M].西安:西安交通大学出版,1996:168-206.
[25]陈寅,林良明,颜国正.D-S证据推理在信息融合应用中的存在问题及改进[J].系统工程与电子,2000,22(11):69-71.
[26]侯彩虹.基于Web的纺织品智能感官评估系统[硕士论文].上海:东华大学,2003.
[27]Tipping,M.E.and C.M.Bishop.Mixtures of probabilistic principal component analyzers[J].Neural Computation,1999,11(2):443-482.
[28]Tipping,M.E,and C.M.Bishop.Probabilistic principal component analysis[J].Journal of the Royal Statistical Society,1999,21(3):611-622.
[29]黄宁.关于主成分分析应用的思考[[J].数理统计与管理,1999,18(5):44-52.
[30]叶双峰.关于主成分分析做综合评价的改进[[J].数理统计与管理,2001,20(2):52-61.
[31]王引斌.测定核心期刊的新方法-主成分分析法[[J].情报学报,1998,17(5):13-16
[32]JianYu.General C-Means Clustering Model[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(8):1197-1211
[33]T.Kanungo,Mount,D.M.Netanyahu,N.S.Piatko,C.D.Silverman,R.Wu,A.Y.An Efficient K-means Clustering Algorithm:Analysis and Implementation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(7):881-892.
[34]Nikhil R.Pal,James C.Bezdek,and Richard J.Hath Learning and the Fuzzy c-Means Clustering Algorithms[J]..Sequential Competitive Neural Networks,1996,9(5):787-796
[35]XU Yang,ZENG Xian-yi,RUAN Da.On Normalization of Comprehensive Evaluation System with Complex Uncertain Information.Journal of Donghua University(Eng,Dd.),2004.21(3):30-34.
[36]Cochran WG.and Cox GM,Experimental designs,2nd ed.[M],John Wiley &Sons,New York,1957:15-35
[37]Pukelsheim F,Optimal design of experiments[M],John Wiley & Sons,New York, USA,1993
[38]Peace GS.,Taguchi methods:a hands-on approach,Addison-Wesley Publishing Company,Reading,MA,USA,1993.
[39]Herbert Stone and Joel Sidel,Sensory Evaluation Practices(Third Edition)[M],Tragon Corporation,Redwood City,CA,USA.2003
[40]Herrera F.and Martizez L.,A model based on linguistic 2-tuples for dealing with multigranular hierarchical linguistic contexts in multi-expert decision-making,IEEE Tranc.Systems,Man,and Cybernetics-Part B:Cybernetics,2001,31(2):227-234
[41]周斌,基于高级计算技术的客户行为刻画的工业产品感官评估研究[博士论文].上海:东华大学,2006
[42]潘泉,于昕等.信息融合理论的基本方法与进展[J].自动化学报.2003,29(4):599-615.
[43]曾庆茂.基于神经网络和模糊推理的信息融合技术[硕士论文].西安:西安科技大学.2005
[44]Kline P,An easy guide to factor analysis,Thousand Oaks(CA):Sage Publications.1994
[45]Mulaik S.A.,The foundations of factor analysis,New York:McGraw-Hill.1972
[46]许丽佳.D-S理论在信息融合中的改进[J].系统工程与电子技术.2004,26(6):717-720.
[47]王宪玉.市场调查与预测[M],北京:中国商业出版社,1996.
[48]Williams J.S.,factor analysis basic theory,Application mathematics andcomputing mathematics,1982(4):41-126.
[49]陈希孺,王松桂.近代回归分析[M],合肥:安徽教育出版社,1987.
[50]J.S.Williams,因子分析理论的综合基础[J],应用数学与计算数学,1982,4:41-126.
[51]Abdi H.,Factor Rotations in Factor Analyses,In M.Lewis-Beck,A.Bryman,T.Futing(Eds):Encyclopedia for research methods for the social sciences,Thousand Oaks(CA):2003:978-982.
[52]Norusis Marija J,SPSS 13.0 Statistical Procedures Companion,Chicago:SPSS,Inc.2005
[53]肖春景,张敏.基于减法聚类与模糊C-均值的模糊聚类的研究.计算机工程.2005,31(增刊):135-137.
[54]Bin Zhou,Xianyi Zeng,Ludovic Koehl,Yongsheng Ding,A 2-level Model for Description of Fabric Hand Sensory Evaluation,shanghai,DongHua university,2004.
[55]周明,孙树栋.遗传算法原理及应用[M].北京:国防工业出版社,1999:13-16
[56]张晓绩,戴冠中,徐乃平.一种新的优化搜索算法-遗传算法[J].控制理论与应用,1995,12(3):265-273
[57]拉夫莱斯等.食品感官评价原理与技术[M].北京:中国轻工业出版社,2001
[2]Dijksterhuis G B.Multivariate Data Analysis in Sensory and Consumer Science[M].Food & Nutrition Press Inc,Trumbull,Connecticut USA,1997.
[3]Ruan Da,The Role of Computational Intelligence in Sensory Evaluation[J].Journal of Donghua University(Eng,Dd.),2004.21(3):21-25.
[4]刘思峰,党耀国,方志耕等.灰色系统理论及其应用[M].北京:科学出版社,2004
[5]Ruan Da,ZENG Xian-y.iIntelligent Sensory Evaluation.Belgium[J].ENSAIT Textile Institute.2003:1-10.
[6]袁红春,钱权,丁德贵.神经网络技术在茶叶感官品质评定中的应用[J].中国茶叶,1999(1):14-15.
[7]周跃斌,尚本清,邓克尼.名优绿茶质量评定方法的研究[J].湖南农业大学学报,1997,23(2):145-149.
[8]赵玉香.感官分析机理与茶叶感官审评条件[J].中国茶叶加工,2001,4:41-42.
[9]覃金钱.感官分析在造纸行业中的应用[J].标准计量与质量,1997(6):37-38.
[10]王光华.感官分析在肉制品加工中的应用[J].肉类工业,1994(12):32-36.
[11]张爱霞,陆淳等.感官分析技术在食品工业中的应用[J].中国乳品工业,2005,33(3):39-40.
[12]张爱霞,邓宏斌等.感官分析技术及在食品工业中的应用[J].乳业科学与技术,2004(3):113-114.
[13]雍克岚,吕敬慈.烟草三维荧光特征和品质评吸的关系[J].化学研究与应用,1998,10(5):494-497.
[14]肖绚等.论陶艺创作中的符号学[J].中国陶瓷,2005(204):32-35.
[15]胡鸿豪,陈刚等.陶瓷智能感官评估平台的研究[J].中国陶瓷,2006(3):36-38
[16]白雪梅等.对主成分分析综合评价方法若干问题的探讨[J].统计学-经济数学方法,1996(1):47-51
[17]Pan N,Zeronian S H,Ryu H-S.An alternative approach to the objective of fabrics[J].Textile Research Journal,1993,63(1):33-43.
[18]Pan N,Yen K C,Zhao S J,et al.A new approach to the objectivemeasurement of fabrics handle from mechanics properties-Part Ⅰ[J].Textile Research Journal,1988,58(8):438-443.
[19]周林峰.基于智能感官评估的服装供应链信息流反馈系统的研究[硕士论文].上海:东华大学,2005
[20]GB/T3532-1995.日用瓷器[S].北京:中国标准出版社,1995
[21]Azeem MF,Hanmandlu M.Generaliantion of adaptive Neuro-fuzzy inference systems[J].IEEE Transactions On Neural Networks,2000
[22]李宏,徐晖,孙仲康.一种具有信息融合功能的模糊推理方法[J].系统工程与电子技术,1998(1):52-54.
[23]罗志增,蒋静坪.基于D-S理论的多信息融合方法及应用[J].电子学报,1999,27(9):100-102.
[24]张文修,梁怡.不确定性推理原理[M].西安:西安交通大学出版,1996:168-206.
[25]陈寅,林良明,颜国正.D-S证据推理在信息融合应用中的存在问题及改进[J].系统工程与电子,2000,22(11):69-71.
[26]侯彩虹.基于Web的纺织品智能感官评估系统[硕士论文].上海:东华大学,2003.
[27]Tipping,M.E.and C.M.Bishop.Mixtures of probabilistic principal component analyzers[J].Neural Computation,1999,11(2):443-482.
[28]Tipping,M.E,and C.M.Bishop.Probabilistic principal component analysis[J].Journal of the Royal Statistical Society,1999,21(3):611-622.
[29]黄宁.关于主成分分析应用的思考[[J].数理统计与管理,1999,18(5):44-52.
[30]叶双峰.关于主成分分析做综合评价的改进[[J].数理统计与管理,2001,20(2):52-61.
[31]王引斌.测定核心期刊的新方法-主成分分析法[[J].情报学报,1998,17(5):13-16
[32]JianYu.General C-Means Clustering Model[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(8):1197-1211
[33]T.Kanungo,Mount,D.M.Netanyahu,N.S.Piatko,C.D.Silverman,R.Wu,A.Y.An Efficient K-means Clustering Algorithm:Analysis and Implementation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(7):881-892.
[34]Nikhil R.Pal,James C.Bezdek,and Richard J.Hath Learning and the Fuzzy c-Means Clustering Algorithms[J]..Sequential Competitive Neural Networks,1996,9(5):787-796
[35]XU Yang,ZENG Xian-yi,RUAN Da.On Normalization of Comprehensive Evaluation System with Complex Uncertain Information.Journal of Donghua University(Eng,Dd.),2004.21(3):30-34.
[36]Cochran WG.and Cox GM,Experimental designs,2nd ed.[M],John Wiley &Sons,New York,1957:15-35
[37]Pukelsheim F,Optimal design of experiments[M],John Wiley & Sons,New York, USA,1993
[38]Peace GS.,Taguchi methods:a hands-on approach,Addison-Wesley Publishing Company,Reading,MA,USA,1993.
[39]Herbert Stone and Joel Sidel,Sensory Evaluation Practices(Third Edition)[M],Tragon Corporation,Redwood City,CA,USA.2003
[40]Herrera F.and Martizez L.,A model based on linguistic 2-tuples for dealing with multigranular hierarchical linguistic contexts in multi-expert decision-making,IEEE Tranc.Systems,Man,and Cybernetics-Part B:Cybernetics,2001,31(2):227-234
[41]周斌,基于高级计算技术的客户行为刻画的工业产品感官评估研究[博士论文].上海:东华大学,2006
[42]潘泉,于昕等.信息融合理论的基本方法与进展[J].自动化学报.2003,29(4):599-615.
[43]曾庆茂.基于神经网络和模糊推理的信息融合技术[硕士论文].西安:西安科技大学.2005
[44]Kline P,An easy guide to factor analysis,Thousand Oaks(CA):Sage Publications.1994
[45]Mulaik S.A.,The foundations of factor analysis,New York:McGraw-Hill.1972
[46]许丽佳.D-S理论在信息融合中的改进[J].系统工程与电子技术.2004,26(6):717-720.
[47]王宪玉.市场调查与预测[M],北京:中国商业出版社,1996.
[48]Williams J.S.,factor analysis basic theory,Application mathematics andcomputing mathematics,1982(4):41-126.
[49]陈希孺,王松桂.近代回归分析[M],合肥:安徽教育出版社,1987.
[50]J.S.Williams,因子分析理论的综合基础[J],应用数学与计算数学,1982,4:41-126.
[51]Abdi H.,Factor Rotations in Factor Analyses,In M.Lewis-Beck,A.Bryman,T.Futing(Eds):Encyclopedia for research methods for the social sciences,Thousand Oaks(CA):2003:978-982.
[52]Norusis Marija J,SPSS 13.0 Statistical Procedures Companion,Chicago:SPSS,Inc.2005
[53]肖春景,张敏.基于减法聚类与模糊C-均值的模糊聚类的研究.计算机工程.2005,31(增刊):135-137.
[54]Bin Zhou,Xianyi Zeng,Ludovic Koehl,Yongsheng Ding,A 2-level Model for Description of Fabric Hand Sensory Evaluation,shanghai,DongHua university,2004.
[55]周明,孙树栋.遗传算法原理及应用[M].北京:国防工业出版社,1999:13-16
[56]张晓绩,戴冠中,徐乃平.一种新的优化搜索算法-遗传算法[J].控制理论与应用,1995,12(3):265-273
[57]拉夫莱斯等.食品感官评价原理与技术[M].北京:中国轻工业出版社,2001