Pub Date : 2008-09-01DOI: 10.1109/UKRICIS.2008.4798937
Vlado Menkovski, I. Christou, S. Efremidis
Classifier ensembles have emerged in recent years as a promising research area for boosting pattern recognition systems' performance. We present a new base classifier that utilizes oblique decision tree technology based on support vector machines for the construction of oblique (non-axis parallel) tests on the nodes of the decision tree inducted. We describe a number of heuristic techniques for enhancing the tree construction process by better estimation of the gain obtained by an oblique split at any tree node. We then show how embedding the new classifier in an ensemble of classifiers using the classical Hedge(beta) algorithm boosts performance of the system. Testing 10-fold cross validation on UCI machine learning repository data sets shows that the new hybrid classifiers outperforms on average by more than 2.1% both the WEKA implementation of C4.5 (J48) and the SMO implementation of SVM in WEKA. The application of the particular ensemble algorithm is an excellent fit for online-learning applications where one seeks to improve performance of self-healing dependable computing systems based on reconfiguration by gradually and adaptively learning what constitutes good system configurations.
{"title":"Oblique Decision Trees using embedded Support Vector Machines in classifier ensembles","authors":"Vlado Menkovski, I. Christou, S. Efremidis","doi":"10.1109/UKRICIS.2008.4798937","DOIUrl":"https://doi.org/10.1109/UKRICIS.2008.4798937","url":null,"abstract":"Classifier ensembles have emerged in recent years as a promising research area for boosting pattern recognition systems' performance. We present a new base classifier that utilizes oblique decision tree technology based on support vector machines for the construction of oblique (non-axis parallel) tests on the nodes of the decision tree inducted. We describe a number of heuristic techniques for enhancing the tree construction process by better estimation of the gain obtained by an oblique split at any tree node. We then show how embedding the new classifier in an ensemble of classifiers using the classical Hedge(beta) algorithm boosts performance of the system. Testing 10-fold cross validation on UCI machine learning repository data sets shows that the new hybrid classifiers outperforms on average by more than 2.1% both the WEKA implementation of C4.5 (J48) and the SMO implementation of SVM in WEKA. The application of the particular ensemble algorithm is an excellent fit for online-learning applications where one seeks to improve performance of self-healing dependable computing systems based on reconfiguration by gradually and adaptively learning what constitutes good system configurations.","PeriodicalId":311925,"journal":{"name":"2008 7th IEEE International Conference on Cybernetic Intelligent Systems","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125627474","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2008-09-01DOI: 10.1109/UKRICIS.2008.4798974
Rongqing Yi, Wenhui Li, X. Cheng
To solve the constraint multi-solution problem, the constraints are separated to two sets, the original constraint set and the additional constraint set. First, the solver finds out multiple solutions. Then genetic algorithm and ant algorithm are combined in the process of searching optimal solution. We adopt genetic algorithm in the former process to produce the initiatory distribution of information elements, and then ant algorithm in the latter process. The random colony is adopted in genetic algorithm, which can not only accelerate the convergence process of ant algorithm but also avoid the local best solution. The heuristic searching algorithm maximizes the fitness of the additional constraint set, thereby reaches the final result that can satisfy the user's expectation.
{"title":"Selecting solutions problem in geometric constraint solving","authors":"Rongqing Yi, Wenhui Li, X. Cheng","doi":"10.1109/UKRICIS.2008.4798974","DOIUrl":"https://doi.org/10.1109/UKRICIS.2008.4798974","url":null,"abstract":"To solve the constraint multi-solution problem, the constraints are separated to two sets, the original constraint set and the additional constraint set. First, the solver finds out multiple solutions. Then genetic algorithm and ant algorithm are combined in the process of searching optimal solution. We adopt genetic algorithm in the former process to produce the initiatory distribution of information elements, and then ant algorithm in the latter process. The random colony is adopted in genetic algorithm, which can not only accelerate the convergence process of ant algorithm but also avoid the local best solution. The heuristic searching algorithm maximizes the fitness of the additional constraint set, thereby reaches the final result that can satisfy the user's expectation.","PeriodicalId":311925,"journal":{"name":"2008 7th IEEE International Conference on Cybernetic Intelligent Systems","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131729244","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2008-09-01DOI: 10.1109/UKRICIS.2008.4798961
Yunqiang Shi, Xianchuan Yu, Xiaochun Cheng, Di Peng
Blind Signal Separation (BSS) and Independent Component Analysis (ICA) are emerging techniques of array processing and data analysis that aim to recover unobserved signals or ldquosourcesrdquo from observed mixtures. When the number of independent components is not very many, using ICA algorithm to separate mixed images, the separated images will have a very strong image artifact noise which will affect the separation efficiency of ICA. In this paper, data fusion technology will be carried out with ICA, firstly, the algorithm used Weiner Filter to preprocess data, and then center the data to make its mean zero; whiten data to make it decorrelation; secondly, use a fast fixed-point algorithm based on negentropy to separate the data, finally, use data fusion based on wavelet algorithm. The results showed that the images after treated by this algorithm become clearer, easier identification, noise significant inhibitory effect. Comparing with the images Peak Signal Noise Ratio (PSNR) before and after the fusion, it proves that the algorithm is feasibility and effectiveness.
{"title":"A new BSS algorithm based on the data fusion and ICA","authors":"Yunqiang Shi, Xianchuan Yu, Xiaochun Cheng, Di Peng","doi":"10.1109/UKRICIS.2008.4798961","DOIUrl":"https://doi.org/10.1109/UKRICIS.2008.4798961","url":null,"abstract":"Blind Signal Separation (BSS) and Independent Component Analysis (ICA) are emerging techniques of array processing and data analysis that aim to recover unobserved signals or ldquosourcesrdquo from observed mixtures. When the number of independent components is not very many, using ICA algorithm to separate mixed images, the separated images will have a very strong image artifact noise which will affect the separation efficiency of ICA. In this paper, data fusion technology will be carried out with ICA, firstly, the algorithm used Weiner Filter to preprocess data, and then center the data to make its mean zero; whiten data to make it decorrelation; secondly, use a fast fixed-point algorithm based on negentropy to separate the data, finally, use data fusion based on wavelet algorithm. The results showed that the images after treated by this algorithm become clearer, easier identification, noise significant inhibitory effect. Comparing with the images Peak Signal Noise Ratio (PSNR) before and after the fusion, it proves that the algorithm is feasibility and effectiveness.","PeriodicalId":311925,"journal":{"name":"2008 7th IEEE International Conference on Cybernetic Intelligent Systems","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123412838","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2008-09-01DOI: 10.1109/UKRICIS.2008.4798949
S. Anishenko, V. Osinov, D. Shaposhnikov, L. Podladchikova, R. Comley, K. Sukholentsev, X. Gao
A motion correction system for brain tomography is presented, which employs two calibrated video cameras and three vision models (CIECAM'97, RETINA, and BMV). The system is evaluated on the pictures monitoring a subjects head while simulating PET scanning (n=12) and on the face images of subjects with different skin colours (n=31). The results on 2D images with known moving parameters show that motion parameters of rotation and translation along X, Y, and Z directions can be obtained very accurately via the method of facial landmark detection. A feasibility study of the developed algorithms is evaluated by a computer simulation of brain scanning procedures.
{"title":"A motion correction system for brain tomography based on biologically motivated models","authors":"S. Anishenko, V. Osinov, D. Shaposhnikov, L. Podladchikova, R. Comley, K. Sukholentsev, X. Gao","doi":"10.1109/UKRICIS.2008.4798949","DOIUrl":"https://doi.org/10.1109/UKRICIS.2008.4798949","url":null,"abstract":"A motion correction system for brain tomography is presented, which employs two calibrated video cameras and three vision models (CIECAM'97, RETINA, and BMV). The system is evaluated on the pictures monitoring a subjects head while simulating PET scanning (n=12) and on the face images of subjects with different skin colours (n=31). The results on 2D images with known moving parameters show that motion parameters of rotation and translation along X, Y, and Z directions can be obtained very accurately via the method of facial landmark detection. A feasibility study of the developed algorithms is evaluated by a computer simulation of brain scanning procedures.","PeriodicalId":311925,"journal":{"name":"2008 7th IEEE International Conference on Cybernetic Intelligent Systems","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128651655","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2008-09-01DOI: 10.1109/UKRICIS.2008.4798925
Md. Zahidul Islam, Chi-Min Oh, Jeong-Seok Yang, Chil-Woo Lee
Robust and real-time object tracking of any objects is a challenging task. Particle filtering has been proven very successful for non-gaussian and non-linear estimation problems. This paper describes a new approach to improve the moving object tracking system with particle filter using shape similarity. The shape similarity between a template and estimated regions in the video sequences can be measured by their normalized cross-correlation of distance transformation. Our observation model of the particle filter is based on shape from distance transformed (DT) edge features. Template is created instantly by selecting any object in a video scene and updated in every frame. Experimental results have been offered to show the effectiveness of the proposed method.
{"title":"DT template based moving object tracking with shape information by particle filter","authors":"Md. Zahidul Islam, Chi-Min Oh, Jeong-Seok Yang, Chil-Woo Lee","doi":"10.1109/UKRICIS.2008.4798925","DOIUrl":"https://doi.org/10.1109/UKRICIS.2008.4798925","url":null,"abstract":"Robust and real-time object tracking of any objects is a challenging task. Particle filtering has been proven very successful for non-gaussian and non-linear estimation problems. This paper describes a new approach to improve the moving object tracking system with particle filter using shape similarity. The shape similarity between a template and estimated regions in the video sequences can be measured by their normalized cross-correlation of distance transformation. Our observation model of the particle filter is based on shape from distance transformed (DT) edge features. Template is created instantly by selecting any object in a video scene and updated in every frame. Experimental results have been offered to show the effectiveness of the proposed method.","PeriodicalId":311925,"journal":{"name":"2008 7th IEEE International Conference on Cybernetic Intelligent Systems","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126674373","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2008-09-01DOI: 10.1109/UKRICIS.2008.4798946
P. Passmore, C. Huyck
Hebb considered the cell assembly to represent a concept in the brain and thus to be an underlying construct of human thought. He proposed that the cell assembly is a connected group of neurons whose pattern of firing is such that a reverberatory activity persists after the removal of a stimulus. Once a cell assembly is activated something must eventually cause it to decay. Clearly thoughts have to be extinguished to make way for others, the question is how. Various suggestions have been made concerning mechanisms that could cause an assembly to decay in the long term including inhibition by other assemblies and passive fatigue. In this paper two classes of models are used to implement this decay, the first is based on building cell assemblies with specific weights and connections that have a linear decay. The second class is based on manipulating variables within a cell assembly model, creating long term fatigue or activation decay. This class of models may be more biologically plausible than the first, and can produce the expected temporal dynamics in the presence of an ambiguous stimulus. However neither class can yet produce the correct prolongation of activation when the stimulus is re-presented.
{"title":"Models of cell assembly decay","authors":"P. Passmore, C. Huyck","doi":"10.1109/UKRICIS.2008.4798946","DOIUrl":"https://doi.org/10.1109/UKRICIS.2008.4798946","url":null,"abstract":"Hebb considered the cell assembly to represent a concept in the brain and thus to be an underlying construct of human thought. He proposed that the cell assembly is a connected group of neurons whose pattern of firing is such that a reverberatory activity persists after the removal of a stimulus. Once a cell assembly is activated something must eventually cause it to decay. Clearly thoughts have to be extinguished to make way for others, the question is how. Various suggestions have been made concerning mechanisms that could cause an assembly to decay in the long term including inhibition by other assemblies and passive fatigue. In this paper two classes of models are used to implement this decay, the first is based on building cell assemblies with specific weights and connections that have a linear decay. The second class is based on manipulating variables within a cell assembly model, creating long term fatigue or activation decay. This class of models may be more biologically plausible than the first, and can produce the expected temporal dynamics in the presence of an ambiguous stimulus. However neither class can yet produce the correct prolongation of activation when the stimulus is re-presented.","PeriodicalId":311925,"journal":{"name":"2008 7th IEEE International Conference on Cybernetic Intelligent Systems","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127334462","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2008-09-01DOI: 10.1109/UKRICIS.2008.4798951
P. van der Putten, L. Meng, J. Kok
In this paper we present an approach for bench-marking and profiling novel classification algorithms. We apply it to AIRS, an artificial immune system algorithm inspired by how the natural immune system recognizes and remembers intruders. We provide basic benchmarking results for AIRS, to our knowledge the first such test under standardised conditions. We also investigate how data set properties (data set size) relate to AIRS performance, and what other algorithms produce similar patterns over over- and underperformance on specific data sets. We present three methods for computing algorithm similarity that may be useful for profiling novel algorithms in general.
{"title":"Profiling novel classification algorithms Artificial Immune Systems","authors":"P. van der Putten, L. Meng, J. Kok","doi":"10.1109/UKRICIS.2008.4798951","DOIUrl":"https://doi.org/10.1109/UKRICIS.2008.4798951","url":null,"abstract":"In this paper we present an approach for bench-marking and profiling novel classification algorithms. We apply it to AIRS, an artificial immune system algorithm inspired by how the natural immune system recognizes and remembers intruders. We provide basic benchmarking results for AIRS, to our knowledge the first such test under standardised conditions. We also investigate how data set properties (data set size) relate to AIRS performance, and what other algorithms produce similar patterns over over- and underperformance on specific data sets. We present three methods for computing algorithm similarity that may be useful for profiling novel algorithms in general.","PeriodicalId":311925,"journal":{"name":"2008 7th IEEE International Conference on Cybernetic Intelligent Systems","volume":"137 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124470392","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2008-09-01DOI: 10.1109/UKRICIS.2008.4798972
Zhong Wang, Yanling Hao, Z. Xiong, F. Sun
In this paper, the radial basis vector (RBV) is proposed to describe the descriptor set of an image. And the shared nearest neighbor clustering kernel (SNNCK) technique is proposed to match RBV pairs. SNNCK is based on the charge attractive model, which will make the unequal-dimensional data sets clustering naturally. Thus, this novel algorithm is able to match the unequal-dimensional data sets when the number of descriptors of two images are unequal. It also can automatically extract the repetition pattern of the reference date set, which is helpful to avoid the wrong matching. Experimental results are also provided, and these results demonstrate superior performances of SNNCK algorithm by using the feature point sets with strong disturbs.
{"title":"SNN clustering kernel technique for content-based scene matching","authors":"Zhong Wang, Yanling Hao, Z. Xiong, F. Sun","doi":"10.1109/UKRICIS.2008.4798972","DOIUrl":"https://doi.org/10.1109/UKRICIS.2008.4798972","url":null,"abstract":"In this paper, the radial basis vector (RBV) is proposed to describe the descriptor set of an image. And the shared nearest neighbor clustering kernel (SNNCK) technique is proposed to match RBV pairs. SNNCK is based on the charge attractive model, which will make the unequal-dimensional data sets clustering naturally. Thus, this novel algorithm is able to match the unequal-dimensional data sets when the number of descriptors of two images are unequal. It also can automatically extract the repetition pattern of the reference date set, which is helpful to avoid the wrong matching. Experimental results are also provided, and these results demonstrate superior performances of SNNCK algorithm by using the feature point sets with strong disturbs.","PeriodicalId":311925,"journal":{"name":"2008 7th IEEE International Conference on Cybernetic Intelligent Systems","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125507977","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2008-09-01DOI: 10.1109/UKRICIS.2008.4798920
J. George, K. Rajeev
Support vector regression using a hybrid wavelet support vector kernel is presented in this paper. A hybrid wavelet kernel construction for support vector machine is introduced. The construction involves a multi-dimensional sinc wavelet function together with one of the conventional kernel functions. We show that the hybrid kernel is an admissible support vector (SV) kernel. The hybrid kernels thus constructed are used for the function approximation problem. The experimental results show that the hybrid kernels provide better function approximation in the mapped feature space compared to conventional kernels.
{"title":"Hybrid wavelet Support Vector Regression","authors":"J. George, K. Rajeev","doi":"10.1109/UKRICIS.2008.4798920","DOIUrl":"https://doi.org/10.1109/UKRICIS.2008.4798920","url":null,"abstract":"Support vector regression using a hybrid wavelet support vector kernel is presented in this paper. A hybrid wavelet kernel construction for support vector machine is introduced. The construction involves a multi-dimensional sinc wavelet function together with one of the conventional kernel functions. We show that the hybrid kernel is an admissible support vector (SV) kernel. The hybrid kernels thus constructed are used for the function approximation problem. The experimental results show that the hybrid kernels provide better function approximation in the mapped feature space compared to conventional kernels.","PeriodicalId":311925,"journal":{"name":"2008 7th IEEE International Conference on Cybernetic Intelligent Systems","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121954890","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2008-09-01DOI: 10.1109/UKRICIS.2008.4798959
Kyarash Shahriari, S. Tarasiewicz
In this paper, a method is proposed to predict n-step response of linear time-varying systems (LTVS). To take into account the modeling error in set-membership approach, adjustable parameters are characterized by intervals. To increase modeling precision, piecewise parameters are used and the validity domain of any set of parameters is determined in input-output space. The output of a model whose parameters are characterized by interval is an interval at any instant. The set of outputs provides a tube for all possible systems responses which is called wrapping envelope. To calculate wrapping envelope, a simulator is proposed who determines at any instant, the valid set of parameters, injects it into the model and calculates interval of systems response. The set of intervals will then determines the wrapping envelope.
{"title":"Response prediction of complex dynamic systems using piecewise parameters","authors":"Kyarash Shahriari, S. Tarasiewicz","doi":"10.1109/UKRICIS.2008.4798959","DOIUrl":"https://doi.org/10.1109/UKRICIS.2008.4798959","url":null,"abstract":"In this paper, a method is proposed to predict n-step response of linear time-varying systems (LTVS). To take into account the modeling error in set-membership approach, adjustable parameters are characterized by intervals. To increase modeling precision, piecewise parameters are used and the validity domain of any set of parameters is determined in input-output space. The output of a model whose parameters are characterized by interval is an interval at any instant. The set of outputs provides a tube for all possible systems responses which is called wrapping envelope. To calculate wrapping envelope, a simulator is proposed who determines at any instant, the valid set of parameters, injects it into the model and calculates interval of systems response. The set of intervals will then determines the wrapping envelope.","PeriodicalId":311925,"journal":{"name":"2008 7th IEEE International Conference on Cybernetic Intelligent Systems","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122051172","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：