改进的ReliefF算法在哈萨克斯拉夫文识别中的应用
|
摘要
针对模式识别领域中的文字识别问题,为达到较高的识别效率,去除低冗余,选择并提取高效特征,提出以印刷体哈萨克斯拉夫字符、符号及数字为研究目标的基于改进的ReliefF特征选择算法,用于哈萨克斯拉夫文字的识别;对传统的ReliefF算法的特征权重做进一步改进,提出Re-reliefF算法,对文字识别的特征选择及特征提取部分进行优化。将基于改进的ReliefF特征选择算法用于哈萨克斯拉夫文识别问题,将该算法的识别结果与传统ReliefF特征选择算法进行比较,比较结果表明,该算法在哈萨克斯拉夫文文字识别过程中效果良好,具有运行速度快与运行时间降低等特点。
As to character recognition problem in the pattern recognition field,to achieve high efficient feature selection and low redundancy,as well as improve the recognition efficiency,an improved ReliefF feature selection algorithm based Kazakhstan printed Cyrillic character recognition method,which selected Kazakh Cyrillic characters,numbers,and punctuation as main object for the study,was proposed.A further study on the features of traditional ReliefF feature selection algorithm was carried out,and the feature weight calculation was improved to achieve Re-reliefF feature selection algorithm,and the feature part of the character recognition system was optimized.The improved Re-reliefF feature selection method was applied in Kazakh Cyrillic identification system,and the recognition result of this algorithm was compared with the traditional ReliefF feature selection algorithm.Results demonstrate that this method performs better in terms of running speed and so on in Kazakh Cyrillic characters recognition system.
As to character recognition problem in the pattern recognition field,to achieve high efficient feature selection and low redundancy,as well as improve the recognition efficiency,an improved ReliefF feature selection algorithm based Kazakhstan printed Cyrillic character recognition method,which selected Kazakh Cyrillic characters,numbers,and punctuation as main object for the study,was proposed.A further study on the features of traditional ReliefF feature selection algorithm was carried out,and the feature weight calculation was improved to achieve Re-reliefF feature selection algorithm,and the feature part of the character recognition system was optimized.The improved Re-reliefF feature selection method was applied in Kazakh Cyrillic identification system,and the recognition result of this algorithm was compared with the traditional ReliefF feature selection algorithm.Results demonstrate that this method performs better in terms of running speed and so on in Kazakh Cyrillic characters recognition system.
引文
[1]DUAN Mingxiu,HE Yingsheng.LVQ neural network based handwritten characters recognition[J].Journal of Jishou University(Natural Science Edition),2010,10(2):41-43(in Chinese).[段明秀,何迎生.基于LVQ神经网络的手写字母识别[J].吉首大学学报(自然科学版),2010,10(2):41-43.]
[2]YIN Xuan,WANG Sheng.Researches on letter distinguish based on discrete Hopfield neural network[J].Computer and Digital Engineering,2011,12(1):34-37(in Chinese).[殷璇,王生.基于离散Hopfield神经网络的字母识别研究[J].计算机与数字工程,2011,12(1):34-37.]
[3]CHEN Qing,YUAN Baoshe,LI Xiao,et al.Printed Uyghur character recognition based on template matching[J].Computer Technology and Development,2012,22(4):119-122(in Chinese).[陈卿,袁保社,李晓,等.基于模板匹配的印刷维吾尔文字符识别研究[J].计算机技术与发展,2012,22(4):119-122.]
[4]Adilijiang·Muhetaer,Yilihamu·Yaermaimaiti,Anwaier·Kasimu.Research on design of Uighur character recognition system[J].Technology Wind,2015(8):25-26(in Chinese).[阿迪力江·穆合塔尔,伊力哈木·亚尔买买提,安外尔·喀斯木.维吾尔文字符识别系统的设计研究[J].科技风,2015(8):25-26.]
[5]MA Weizhen.Research on on-line handwritten recognition of Tibetan characters[D].Lhasa:Tibet University,2013(in Chinese).[马伟珍.联机手写藏文识别系统研究[D].拉萨:西藏大学,2013.]
[6]CHUN Hua.Research of printed Mongolian characters recognition[J].Journal of Inner Mngolia University for Nationalities,2014,29(6):627-628(in Chinese).[春花.印刷体蒙古文文字识别的研究[J].内蒙古民族大学学报:自然科学版,2014,29(6):627-628.]
[7]YAO Xu,WANG Xiaodan,ZHANG Yuxi,et al.Summary of feature selection algorithms[J].Control and Decision,2012,27(2):161-166(in Chinese).[姚旭,王晓丹,张玉玺,等.特征选择方法综述[J].控制与决策,2012,27(2):161-166.]
[8]WU Hongxia,WU Yue,LIU Zongtian,et al.Combined SNP feature selection based on Relief and SVM-REF[J].Journal of Application Research of Computers,2012,29(6):2074-2077(in Chinese).[吴红霞,吴悦,刘宗田,等.基于Relief和SVM-RFE的组合式SNP特征选择[J].计算机应用研究,2012,29(6):2074-2077.]
[9]Zhang N,Zhou Y,Huang T,et al.Discriminating between lysine sumoylation and lysine acetylation using mRMR feature selection and analysis[J].PLoS ONE,2014,9(9):e107464.
[10]JIA Xue.Research on intrusion feature selection algorithms based on ReliefF[D].Urumqi:Xinjiang University,2014(in Chinese).[贾雪.基于ReliefF的入侵特征选择算法研究[D].乌鲁木齐:新疆大学,2014.]
[11]LIU Jinlong,ZHANG Yan,DONG Jun.The study on the conversion rules between character encodings of Slavic-based Kazakh and Arabic-based Kazakh[J].Electronic Test,2014(10):144-146(in Chinese).[刘金龙,张岩,董军.斯拉夫哈萨克文与现行哈萨克文编码字符转换规则的探究[J].电子测试,2014(10):144-146.]
[12]Sahdolla·Mubarak,Gulila·Altenbek.Research on intelligent conversion between Kazakh both text forms[J].Computer Engineering and Applications,2014,50(18):226-229(in Chinese).[萨合多拉·木巴拉克,古丽拉·阿东别克.哈萨克语阿拉伯文与斯拉夫文间的智能转换[J].计算机工程与应用,2014,50(18):226-229.]
[13]YANG Feihu.Research on feature selection approach and its application in network traffic identification[D].Nanjing:Nanjing University of Posts and Telecommunications,2012(in Chinese).[杨飞虎.特征选择算法及其在网络流量识别中的应用研究[D].南京:南京邮电大学,2012.]
[14]Rashedi E,Nezamabadi-Pour H,Saryazdi S.A simultaneous feature adaptation and feature selection method for contentbased image retrieval systems[J].Knowledge-Based Systems,2013,39(2):85-94.
[15]Shang C,Li M,Feng S,et al.Feature selection via maximizing global information gain for text classification[J].Knowledge-Based Systems,2013,54(4):298-309.
[16]Zhang H,Zhang Y,Huang TS.Pose-robust face recognition via sparse representation[J].Pattern Recognition,2013,46(5):1511-1521.
[17]CHEN Donghui.Reasearch of key techniques in fuzzy clustering based on objective function[D].Xi’an:Xi’an Electronic and Science University,2012(in Chinese).[陈东辉.基于目标函数的模糊聚类算法关键技术研究[D].西安:西安电子科技大学,2012.]
[18]HUANG Lili,TANG Jin,SUN Dengdi,et al.Feature selection algorithim based on multi-label ReliefF[J].Journal of Computer Application,2012,32(10):2888-2890(in Chinese).[黄莉莉,汤进,孙登第,等.基于多标签ReliefF的特征选择算法[J].计算机应用,2012,32(10):2888-2890.]
[19]WANG Lu,GONG Guanghong.Multiple features remote sensing image classification based on combining ReliefF and mRMR[J].Chinese Journal of Stereology and Image Analysis,2014(3):250-257(in Chinese).[王露,龚光红.基于ReliefF+mRMR特征降维算法的多特征遥感图像分类[J].中国体视学与图像分析,2014(3):250-257.]
[20]YE Jixiang,WANG Conghui.Application of improvement of F-scroe algorithm in speech emotional recognition[J].Computer Engineering and Applications,2013,49(16):137-141(in Chinese).[叶吉祥,王聪慧.改进的F-score算法在语音情感识别中的应用[J].计算机工程与应用,2013,49(16):137-141.]
[2]YIN Xuan,WANG Sheng.Researches on letter distinguish based on discrete Hopfield neural network[J].Computer and Digital Engineering,2011,12(1):34-37(in Chinese).[殷璇,王生.基于离散Hopfield神经网络的字母识别研究[J].计算机与数字工程,2011,12(1):34-37.]
[3]CHEN Qing,YUAN Baoshe,LI Xiao,et al.Printed Uyghur character recognition based on template matching[J].Computer Technology and Development,2012,22(4):119-122(in Chinese).[陈卿,袁保社,李晓,等.基于模板匹配的印刷维吾尔文字符识别研究[J].计算机技术与发展,2012,22(4):119-122.]
[4]Adilijiang·Muhetaer,Yilihamu·Yaermaimaiti,Anwaier·Kasimu.Research on design of Uighur character recognition system[J].Technology Wind,2015(8):25-26(in Chinese).[阿迪力江·穆合塔尔,伊力哈木·亚尔买买提,安外尔·喀斯木.维吾尔文字符识别系统的设计研究[J].科技风,2015(8):25-26.]
[5]MA Weizhen.Research on on-line handwritten recognition of Tibetan characters[D].Lhasa:Tibet University,2013(in Chinese).[马伟珍.联机手写藏文识别系统研究[D].拉萨:西藏大学,2013.]
[6]CHUN Hua.Research of printed Mongolian characters recognition[J].Journal of Inner Mngolia University for Nationalities,2014,29(6):627-628(in Chinese).[春花.印刷体蒙古文文字识别的研究[J].内蒙古民族大学学报:自然科学版,2014,29(6):627-628.]
[7]YAO Xu,WANG Xiaodan,ZHANG Yuxi,et al.Summary of feature selection algorithms[J].Control and Decision,2012,27(2):161-166(in Chinese).[姚旭,王晓丹,张玉玺,等.特征选择方法综述[J].控制与决策,2012,27(2):161-166.]
[8]WU Hongxia,WU Yue,LIU Zongtian,et al.Combined SNP feature selection based on Relief and SVM-REF[J].Journal of Application Research of Computers,2012,29(6):2074-2077(in Chinese).[吴红霞,吴悦,刘宗田,等.基于Relief和SVM-RFE的组合式SNP特征选择[J].计算机应用研究,2012,29(6):2074-2077.]
[9]Zhang N,Zhou Y,Huang T,et al.Discriminating between lysine sumoylation and lysine acetylation using mRMR feature selection and analysis[J].PLoS ONE,2014,9(9):e107464.
[10]JIA Xue.Research on intrusion feature selection algorithms based on ReliefF[D].Urumqi:Xinjiang University,2014(in Chinese).[贾雪.基于ReliefF的入侵特征选择算法研究[D].乌鲁木齐:新疆大学,2014.]
[11]LIU Jinlong,ZHANG Yan,DONG Jun.The study on the conversion rules between character encodings of Slavic-based Kazakh and Arabic-based Kazakh[J].Electronic Test,2014(10):144-146(in Chinese).[刘金龙,张岩,董军.斯拉夫哈萨克文与现行哈萨克文编码字符转换规则的探究[J].电子测试,2014(10):144-146.]
[12]Sahdolla·Mubarak,Gulila·Altenbek.Research on intelligent conversion between Kazakh both text forms[J].Computer Engineering and Applications,2014,50(18):226-229(in Chinese).[萨合多拉·木巴拉克,古丽拉·阿东别克.哈萨克语阿拉伯文与斯拉夫文间的智能转换[J].计算机工程与应用,2014,50(18):226-229.]
[13]YANG Feihu.Research on feature selection approach and its application in network traffic identification[D].Nanjing:Nanjing University of Posts and Telecommunications,2012(in Chinese).[杨飞虎.特征选择算法及其在网络流量识别中的应用研究[D].南京:南京邮电大学,2012.]
[14]Rashedi E,Nezamabadi-Pour H,Saryazdi S.A simultaneous feature adaptation and feature selection method for contentbased image retrieval systems[J].Knowledge-Based Systems,2013,39(2):85-94.
[15]Shang C,Li M,Feng S,et al.Feature selection via maximizing global information gain for text classification[J].Knowledge-Based Systems,2013,54(4):298-309.
[16]Zhang H,Zhang Y,Huang TS.Pose-robust face recognition via sparse representation[J].Pattern Recognition,2013,46(5):1511-1521.
[17]CHEN Donghui.Reasearch of key techniques in fuzzy clustering based on objective function[D].Xi’an:Xi’an Electronic and Science University,2012(in Chinese).[陈东辉.基于目标函数的模糊聚类算法关键技术研究[D].西安:西安电子科技大学,2012.]
[18]HUANG Lili,TANG Jin,SUN Dengdi,et al.Feature selection algorithim based on multi-label ReliefF[J].Journal of Computer Application,2012,32(10):2888-2890(in Chinese).[黄莉莉,汤进,孙登第,等.基于多标签ReliefF的特征选择算法[J].计算机应用,2012,32(10):2888-2890.]
[19]WANG Lu,GONG Guanghong.Multiple features remote sensing image classification based on combining ReliefF and mRMR[J].Chinese Journal of Stereology and Image Analysis,2014(3):250-257(in Chinese).[王露,龚光红.基于ReliefF+mRMR特征降维算法的多特征遥感图像分类[J].中国体视学与图像分析,2014(3):250-257.]
[20]YE Jixiang,WANG Conghui.Application of improvement of F-scroe algorithm in speech emotional recognition[J].Computer Engineering and Applications,2013,49(16):137-141(in Chinese).[叶吉祥,王聪慧.改进的F-score算法在语音情感识别中的应用[J].计算机工程与应用,2013,49(16):137-141.]