Pub Date : 2017-06-01DOI: 10.1109/WSOM.2017.8020011
Nobuo Matsuda, H. Tokutaka, Hideaki Sato, F. Tajima, Reiji Kawata
This paper describes the filtering effects on classification performance with applying significance degree by SOM to the image analysis using filter bank preprocessing and Subspace Classifier. In our proposed method, a series of analysis concerning accuracy were first conducted in the cases of single filter and filter bank, and examinations on significance degree by SOM were conducted based on green(G) and blue(B) color channels. The difference of the filtering effect between two color channels was compared with using the significance degree. We show that the difference of the filtering effect between two channels can be clarified by using the significance degree by SOM.
{"title":"Applying the significance degree by SOM to image analysis of fundus using the filter bank","authors":"Nobuo Matsuda, H. Tokutaka, Hideaki Sato, F. Tajima, Reiji Kawata","doi":"10.1109/WSOM.2017.8020011","DOIUrl":"https://doi.org/10.1109/WSOM.2017.8020011","url":null,"abstract":"This paper describes the filtering effects on classification performance with applying significance degree by SOM to the image analysis using filter bank preprocessing and Subspace Classifier. In our proposed method, a series of analysis concerning accuracy were first conducted in the cases of single filter and filter bank, and examinations on significance degree by SOM were conducted based on green(G) and blue(B) color channels. The difference of the filtering effect between two color channels was compared with using the significance degree. We show that the difference of the filtering effect between two channels can be clarified by using the significance degree by SOM.","PeriodicalId":130086,"journal":{"name":"2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124211487","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 : 2017-06-01DOI: 10.1109/WSOM.2017.8020029
David Nova, P. Estévez
In this contribution we propose a new regularization method for the Generalized Matrix Learning Vector Quantization classifier. In particular we use a nuclear norm in order to prevent oversimplifying/over-fitting and oscillatory behaviour of the small eigenvalues of the positive semi-definite relevance matrix. The proposed method is compared with two other regularization methods in two artificial data sets and a reallife problem. The results show that the proposed regularization method enhances the generalization ability of GMLVQ. This is reflected in a lower classification error and a better interpretability of the relevance matrix.
{"title":"Spectral regularization in generalized matrix learning vector quantization","authors":"David Nova, P. Estévez","doi":"10.1109/WSOM.2017.8020029","DOIUrl":"https://doi.org/10.1109/WSOM.2017.8020029","url":null,"abstract":"In this contribution we propose a new regularization method for the Generalized Matrix Learning Vector Quantization classifier. In particular we use a nuclear norm in order to prevent oversimplifying/over-fitting and oscillatory behaviour of the small eigenvalues of the positive semi-definite relevance matrix. The proposed method is compared with two other regularization methods in two artificial data sets and a reallife problem. The results show that the proposed regularization method enhances the generalization ability of GMLVQ. This is reflected in a lower classification error and a better interpretability of the relevance matrix.","PeriodicalId":130086,"journal":{"name":"2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127953965","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 : 2017-06-01DOI: 10.1109/WSOM.2017.8020007
J. Deichmann, A. Eshghi, D. Haughton, Mingfei Li
This paper uses SOMbrero visualizations to examine two socio-economic dimensions of European states, generated by a factor analysis of time-series data from 2001–2013. We analyze SOMs for 41 countries with regard to “Old Capital” and “New Capital”, two factors that are generated from 12 variables. SOMbrero reveals evidence of various convergence paths over time for these two factors. This approach also clearly uncovers the differential impacts of the European recession upon clusters of European countries. In conclusion, we demonstrate that SOMs are a useful tool for better understanding European convergence.
{"title":"Using SOMbrero to examine the economic convergence of European countries from 2001–2013","authors":"J. Deichmann, A. Eshghi, D. Haughton, Mingfei Li","doi":"10.1109/WSOM.2017.8020007","DOIUrl":"https://doi.org/10.1109/WSOM.2017.8020007","url":null,"abstract":"This paper uses SOMbrero visualizations to examine two socio-economic dimensions of European states, generated by a factor analysis of time-series data from 2001–2013. We analyze SOMs for 41 countries with regard to “Old Capital” and “New Capital”, two factors that are generated from 12 variables. SOMbrero reveals evidence of various convergence paths over time for these two factors. This approach also clearly uncovers the differential impacts of the European recession upon clusters of European countries. In conclusion, we demonstrate that SOMs are a useful tool for better understanding European convergence.","PeriodicalId":130086,"journal":{"name":"2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)","volume":"242 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131544816","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 : 2017-06-01DOI: 10.1109/WSOM.2017.8020031
Iván Machón-González, Hilario López-García
The dynamic identification of a nonlinear plant is not a trivial issue. The application of a neural gas network that is trained with a supervised batch version of the algorithm can produce identification models in a robust way. In this paper, the neural model identifies each local transfer function demonstrating that the local linear approximation can be done. Moreover, other parameters are analyzed in order to obtain a correct modeling.
{"title":"Nonlinear dynamic identification using supervised neural gas algorithm","authors":"Iván Machón-González, Hilario López-García","doi":"10.1109/WSOM.2017.8020031","DOIUrl":"https://doi.org/10.1109/WSOM.2017.8020031","url":null,"abstract":"The dynamic identification of a nonlinear plant is not a trivial issue. The application of a neural gas network that is trained with a supervised batch version of the algorithm can produce identification models in a robust way. In this paper, the neural model identifies each local transfer function demonstrating that the local linear approximation can be done. Moreover, other parameters are analyzed in order to obtain a correct modeling.","PeriodicalId":130086,"journal":{"name":"2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123733747","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 : 2017-06-01DOI: 10.1109/WSOM.2017.8020022
Ludovic Platon, F. Zehraoui, F. Tahi
We present in this paper a new approach of supervised self organizing map (SOM). We added a supervised perceptron layer to the classical SOM approach. This combination allows the classification of new patterns by taking into account all the map prototypes without changing the SOM organization. We also propose to associate two reject options to our supervised SOM. This allows to improve the results reliability and to discover new classes in applications where some classes are unknown. We obtain two variants of supervised SOM with rejection that have been evaluated on different datasets. The results indicate that our approaches are competitive with most popular supervised leaning algorithms like support vector machines and random forest.
{"title":"Self-organizing maps with supervised layer","authors":"Ludovic Platon, F. Zehraoui, F. Tahi","doi":"10.1109/WSOM.2017.8020022","DOIUrl":"https://doi.org/10.1109/WSOM.2017.8020022","url":null,"abstract":"We present in this paper a new approach of supervised self organizing map (SOM). We added a supervised perceptron layer to the classical SOM approach. This combination allows the classification of new patterns by taking into account all the map prototypes without changing the SOM organization. We also propose to associate two reject options to our supervised SOM. This allows to improve the results reliability and to discover new classes in applications where some classes are unknown. We obtain two variants of supervised SOM with rejection that have been evaluated on different datasets. The results indicate that our approaches are competitive with most popular supervised leaning algorithms like support vector machines and random forest.","PeriodicalId":130086,"journal":{"name":"2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125772944","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 : 2017-06-01DOI: 10.1109/WSOM.2017.8020005
P. Mei, C. C. Carneiro, M. Kuroda, S. Fraser, L. Min, F. Reis
Multiple Sclerosis (MS) is the most prevalent demyelinating disease of the Central Nervous System, being the Relapsing-Remitting (RRMS) its most common subtype. We explored here the viability of use of Self Organizing Maps (SOM) to perform automatic segmentation of MS lesions apart from CNS normal tissue. SOM were able, in most cases, to successfully segment MRIs of patients with RRMS, with the correct separation of normal versus pathological tissue especially in supratentorial acquisitions, although it could not differentiate older from newer lesions.
{"title":"Self-organizing maps as a tool for segmentation of Magnetic Resonance Imaging (MRI) of relapsing-remitting multiple sclerosis","authors":"P. Mei, C. C. Carneiro, M. Kuroda, S. Fraser, L. Min, F. Reis","doi":"10.1109/WSOM.2017.8020005","DOIUrl":"https://doi.org/10.1109/WSOM.2017.8020005","url":null,"abstract":"Multiple Sclerosis (MS) is the most prevalent demyelinating disease of the Central Nervous System, being the Relapsing-Remitting (RRMS) its most common subtype. We explored here the viability of use of Self Organizing Maps (SOM) to perform automatic segmentation of MS lesions apart from CNS normal tissue. SOM were able, in most cases, to successfully segment MRIs of patients with RRMS, with the correct separation of normal versus pathological tissue especially in supratentorial acquisitions, although it could not differentiate older from newer lesions.","PeriodicalId":130086,"journal":{"name":"2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115291218","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 : 2017-06-01DOI: 10.1109/WSOM.2017.8020032
Tina Geweniger, T. Villmann
In this article we propose a relational and a median possibilistic clustering method. Both methods are modifications of Possibilistic Fuzzy C-Means as introduced by Pal et al. [1]. The proposed algorithms are applicable for abstract non-vectorial data objects where only the dissimilarities of the objects are known. For the relational version we assume a Euclidean data embedding. For data where this assumption is not feasible we introduce a median variant restricting prototypes to be data objects themselves.
{"title":"Relational and median variants of Possibilistic Fuzzy C-Means","authors":"Tina Geweniger, T. Villmann","doi":"10.1109/WSOM.2017.8020032","DOIUrl":"https://doi.org/10.1109/WSOM.2017.8020032","url":null,"abstract":"In this article we propose a relational and a median possibilistic clustering method. Both methods are modifications of Possibilistic Fuzzy C-Means as introduced by Pal et al. [1]. The proposed algorithms are applicable for abstract non-vectorial data objects where only the dissimilarities of the objects are known. For the relational version we assume a Euclidean data embedding. For data where this assumption is not feasible we introduce a median variant restricting prototypes to be data objects themselves.","PeriodicalId":130086,"journal":{"name":"2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131105042","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 : 2017-06-01DOI: 10.1109/WSOM.2017.8020033
J. Vergara, P. Estévez
Segment Growing Neural Gas (Segment-GNG) has been recently proposed as a new spatiotemporal quantization method for time series. Unlike traditional quantization algorithms that are prototype-based, Segment-GNG uses segments as basic units of quantization. In this paper we extend the Segment-GNG model in order to deal with time series prediction. First Segment-GNG makes a quantization of the trajectories in the state-space representation of the time series. Then a local prediction model is associated with each segment, which allows us to make predictions. The proposed model is tested with the Mackey-Glass and Lorenz chaotic time series in one-step ahead prediction tasks. The results obtained are competitive with the best results published in the literature.
{"title":"A strategy for time series prediction using Segment Growing Neural Gas","authors":"J. Vergara, P. Estévez","doi":"10.1109/WSOM.2017.8020033","DOIUrl":"https://doi.org/10.1109/WSOM.2017.8020033","url":null,"abstract":"Segment Growing Neural Gas (Segment-GNG) has been recently proposed as a new spatiotemporal quantization method for time series. Unlike traditional quantization algorithms that are prototype-based, Segment-GNG uses segments as basic units of quantization. In this paper we extend the Segment-GNG model in order to deal with time series prediction. First Segment-GNG makes a quantization of the trajectories in the state-space representation of the time series. Then a local prediction model is associated with each segment, which allows us to make predictions. The proposed model is tested with the Mackey-Glass and Lorenz chaotic time series in one-step ahead prediction tasks. The results obtained are competitive with the best results published in the literature.","PeriodicalId":130086,"journal":{"name":"2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127025131","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 : 2017-06-01DOI: 10.1109/WSOM.2017.8020025
C. Mattos, G. Barreto, D. Horstkemper, B. Hellingrath
In this paper we evaluate metaheuristic optimization methods on a partitional clustering task of a real-world supply chain dataset, aiming at customer segmentation. For this purpose, we rely on the automatic clustering framework proposed by Das et al. [1], named henceforth DAK framework, by testing its performance for seven different metaheuristic optimization algorithm, namely: simulated annealing (SA), genetic algorithms (GA), particle swarm optimization (PSO), differential evolution (DE), artificial bee colony (ABC), cuckoo search (CS) and fireworks algorithm (FA). An in-depth analysis of the obtained results is carried out in order to compare the performances of the metaheuristic optimization algorithms under the DAK framework with that of standard (i.e. non-automatic) clustering methodology.
{"title":"Metaheuristic optimization for automatic clustering of customer-oriented supply chain data","authors":"C. Mattos, G. Barreto, D. Horstkemper, B. Hellingrath","doi":"10.1109/WSOM.2017.8020025","DOIUrl":"https://doi.org/10.1109/WSOM.2017.8020025","url":null,"abstract":"In this paper we evaluate metaheuristic optimization methods on a partitional clustering task of a real-world supply chain dataset, aiming at customer segmentation. For this purpose, we rely on the automatic clustering framework proposed by Das et al. [1], named henceforth DAK framework, by testing its performance for seven different metaheuristic optimization algorithm, namely: simulated annealing (SA), genetic algorithms (GA), particle swarm optimization (PSO), differential evolution (DE), artificial bee colony (ABC), cuckoo search (CS) and fireworks algorithm (FA). An in-depth analysis of the obtained results is carried out in order to compare the performances of the metaheuristic optimization algorithms under the DAK framework with that of standard (i.e. non-automatic) clustering methodology.","PeriodicalId":130086,"journal":{"name":"2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)","volume":"87 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130427501","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 : 2017-06-01DOI: 10.1109/WSOM.2017.8020003
M. Kirby, C. Peterson
We extend the self-organizing mapping algorithm to the problem of visualizing data on Grassmann manifolds. In this setting, a collection of k points in n-dimensions is represented by a k-dimensional subspace, e.g., via the singular value or QR-decompositions. Data assembled in this way is challenging to visualize given abstract points on the Grassmannian do not reside in Euclidean space. The extension of the SOM algorithm to this geometric setting only requires that distances between two points can be measured and that any given point can be moved towards a presented pattern. The similarity between two points on the Grassmannian is measured in terms of the principal angles between subspaces, e.g., the chordal distance. Further, we employ a formula for moving one subspace towards another along the shortest path, i.e., the geodesic between two points on the Grassmannian. This enables a faithful implementation of the SOM approach for visualizing data consisting of k-dimensional subspaces of n-dimensional Euclidean space. We illustrate the resulting algorithm on a hyperspectral imaging application.
{"title":"Visualizing data sets on the Grassmannian using self-organizing mappings","authors":"M. Kirby, C. Peterson","doi":"10.1109/WSOM.2017.8020003","DOIUrl":"https://doi.org/10.1109/WSOM.2017.8020003","url":null,"abstract":"We extend the self-organizing mapping algorithm to the problem of visualizing data on Grassmann manifolds. In this setting, a collection of k points in n-dimensions is represented by a k-dimensional subspace, e.g., via the singular value or QR-decompositions. Data assembled in this way is challenging to visualize given abstract points on the Grassmannian do not reside in Euclidean space. The extension of the SOM algorithm to this geometric setting only requires that distances between two points can be measured and that any given point can be moved towards a presented pattern. The similarity between two points on the Grassmannian is measured in terms of the principal angles between subspaces, e.g., the chordal distance. Further, we employ a formula for moving one subspace towards another along the shortest path, i.e., the geodesic between two points on the Grassmannian. This enables a faithful implementation of the SOM approach for visualizing data consisting of k-dimensional subspaces of n-dimensional Euclidean space. We illustrate the resulting algorithm on a hyperspectral imaging application.","PeriodicalId":130086,"journal":{"name":"2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM)","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114206142","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：