2,018 Hits in 6.5 sec

Tackling temporal pattern recognition by vector space embedding

Brian Iwana, Seiichi Uchida, Kaspar Riesen, Volkmar Frinken
2015 2015 13th International Conference on Document Analysis and Recognition (ICDAR)  
To solve this problem, we propose a method of representing the temporal patterns by embedding dynamic time warping (DTW) distance based dissimilarities in vector space.  ...  This paper introduces a novel method of reducing the number of prototype patterns necessary for accurate recognition of temporal patterns.  ...  Fig. 1 . 1 A visualization of embedding an element s into an N -dimensional vector space using pairwise dissimilarities to prototypes.  ... 
doi:10.1109/icdar.2015.7333875 dblp:conf/icdar/IwanaURF15 fatcat:llh4zobtnfasnjzqgshv43mu3u

Adaptive local dissimilarity measures for discriminative dimension reduction of labeled data

Kerstin Bunte, Barbara Hammer, Axel Wismüller, Michael Biehl
2010 Neurocomputing  
In this contribution we use an extension of prototype-based local distance learning, which results in a nonlinear discriminative dissimilarity measure for a given labeled data manifold.  ...  We show the combination of different dimension reduction techniques with a discriminative similarity measure learned by an extension of Learning Vector Quantization (LVQ) and their behavior with different  ...  Step 3: Reconstruct the topology induced by the input data by moving the image vectors in the embedding space χ using the computational scheme of a topology-preserving mapping.  ... 
doi:10.1016/j.neucom.2009.11.017 fatcat:n52n6s3syfamtn4237lbqnsr7i

Prototype Selection for Dissimilarity Representation by a Genetic Algorithm

Yenisel Plasencia Calana, Edel Garcia Reyes, Mauricio Orozco Alzate, Robert P.W. Duin
2010 2010 20th International Conference on Pattern Recognition  
In this paper it is shown that genetic algorithms, previously used for feature selection, may be used for building good dissimilarity spaces as well, especially when small sets of prototypes are needed  ...  The resulting dissimilarity space may be used to train any classifier appropriate for feature spaces. There is, however, a strong need for dimension reduction.  ...  The computation of dissimilarities can be computationally demanding as in case of comparisons of images and graphs, then a reduction of the number of dissimilarities to be measured is of interest.  ... 
doi:10.1109/icpr.2010.52 dblp:conf/icpr/CalanaROD10 fatcat:idcgmlmkh5bkxpgg4libjwltcm

Dissimilarity space reinforced with manifold learning and latent space modeling for improved pattern classification

Azadeh Rezazadeh Hamedani, Mohammad Hossein Moattar, Yahya Forghani
2021 Journal of Big Data  
Dissimilarity space embedding is an approach in which each sample is represented as a vector based on its dissimilarity to some other samples called prototypes.  ...  In order to create the dissimilarity space, each sample is compared only with its prototype set including its k-nearest neighbors on the manifold using the geodesic distance metric.  ...  It should be noted that we prefer to use the initial samples ( X N * D ) before any dimension reduction, so the dissimilarity space is not constructed using manifold embedding Y N * d .  ... 
doi:10.1186/s40537-021-00527-6 fatcat:ee7dao36vfgmxdjz7q4tvtl6yy

Learning vector quantization for (dis-)similarities

Barbara Hammer, Daniela Hofmann, Frank-Michael Schleif, Xibin Zhu
2014 Neurocomputing  
We propose a general framework how the methods can be combined based on the background of a pseudo-Euclidean embedding of the data.  ...  However, popular prototype based classifiers such as learning vector quantization (LVQ) are restricted to vectorial data only.  ...  The authors are responsible for the contents of this publication.  ... 
doi:10.1016/j.neucom.2013.05.054 fatcat:tcpoaidazfhrbj4ky2u25jbgeq

On Using Dimensionality Reduction Schemes to Optimize Dissimilarity-Based Classifiers [chapter]

Sang-Woon Kim, Jian Gao
2008 Lecture Notes in Computer Science  
That is, we prefer not to directly select the representative prototypes from the training samples; rather, we use a dimensionality reduction scheme after computing the dissimilarity matrix with the entire  ...  To avoid these problems, in this paper, we propose an alternative approach where we use all available samples from the training set as prototypes and subsequently apply dimensionality reduction schemes  ...  The main contribution of this paper is to demonstrate that dissimilarity-based classification can be optimized by employing a dimensionality reduction scheme.  ... 
doi:10.1007/978-3-540-85920-8_38 fatcat:4ior3nbivjdk5hc5lkd3dxzw6e

Experimental study on prototype optimisation algorithms for prototype-based classification in vector spaces

M. Lozano, J.M. Sotoca, J.S. Sánchez, F. Pla, E. Pękalska, R.P.W. Duin
2006 Pattern Recognition  
Usually the determination of prototypes is studied in relation with the nearest neighbour rule. We will show that the use of more general dissimilarity-based classifiers can be more beneficial.  ...  These condensing techniques are evaluated on real data, represented in vector spaces, by comparing their resulting reduction rates and classification performance.  ...  DUIN studied applied physics at Delft University of Technology in the Netherlands. In 1978 he received the Ph.D. degree for a thesis on the accuracy of statistical classifiers.  ... 
doi:10.1016/j.patcog.2006.04.005 fatcat:pitfyhastbetbmasatc6ymhzoq

Human Action Recognition in Video by Fusion of Structural and Spatio-temporal Features [chapter]

Ehsan Zare Borzeshi, Oscar Perez Concha, Massimo Piccardi
2012 Lecture Notes in Computer Science  
This paper proposes a novel method based on the fusion of global spatial relationships provided by graph embedding and the local spatio-temporal information of STIP descriptors.  ...  The problem of human action recognition has received increasing attention in recent years for its importance in many applications.  ...  The authors wish to thank the Australian Research Council and its industry partners that have partially supported this work under the Linkage Project funding scheme -grant LP 0990135 "Airport of Future  ... 
doi:10.1007/978-3-642-34166-3_52 fatcat:fdc6a5lclvgfxhkqpsbsg4gqsm

Topographic Mapping of Large Dissimilarity Data Sets

Barbara Hammer, Alexander Hasenfuss
2010 Neural Computation  
In this contribution, we introduce relational topographic maps as an extension of relational clustering algorithms which offer prototype-based representations of dissimilarity data, to incorporate neighborhood  ...  Often, data are available in the form of pairwise distances only, such as e.g. arise from a kernel matrix, a graph, or some general dissimilarity measure.  ...  This cost function is independent of an embedding of data in a vector space.  ... 
doi:10.1162/neco_a_00012 pmid:20569180 fatcat:di5r7afs3jdepjjhilkxh4hk54

Prototype generation on structural data using dissimilarity space representation

Jorge Calvo-Zaragoza, Jose J. Valero-Mas, Juan R. Rico-Juan
2016 Neural computing & applications (Print)  
This work studies the possibility of using Dissimilarity Space (DS) methods as an intermediate process for mapping the initial structural representation to a statistical one, thereby allowing the use of  ...  Data Reduction techniques play a key role in instance-based classification to lower the amount of data to be processed.  ...  Prototype Generation over Structural Data using Dissimilarity Space Representation Current PG algorithms assume that data is defined over a vector space.  ... 
doi:10.1007/s00521-016-2278-8 fatcat:2elbhayncnhgvo4d3ldhm6hudu

Designing Labeled Graph Classifiers by Exploiting the Rényi Entropy of the Dissimilarity Representation

Lorenzo Livi
2017 Entropy  
However, the design of effective learning procedures operating in the space of labeled graphs is still a challenging problem, especially from the computational complexity viewpoint.  ...  In this paper, we present a major improvement of a general-purpose classifier for graphs, which is conceived on an interplay between dissimilarity representation, clustering, information-theoretic techniques  ...  Acronym Full name BSAS Basic sequential algorithmic scheme CBC Clustering-based compression DM Dissimilarity matrix DS Dissimilarity space DV Dissimilarity value IGM Inexact graph matching  ... 
doi:10.3390/e19050216 fatcat:nim5alx6wrblzooxiuln5co7dq

Exact and Inexact Graph Matching: Methodology and Applications [chapter]

Kaspar Riesen, Xiaoyi Jiang, Horst Bunke
2010 Managing and Mining Graph Data  
Graphs provide us with a powerful and flexible representation formalism which can be employed in various fields of intelligent information processing.  ...  The process of evaluating the similarity of graphs is referred to as graph matching. Two approaches to this task exist, viz. exact and inexact graph matching.  ...  In order to overcome this limitation, in [68] , various prototype reduction schemes [3] are adopted for the task of graph embedding.  ... 
doi:10.1007/978-1-4419-6045-0_7 dblp:series/ads/RiesenJB10 fatcat:mr2bpz5xwnbiva5x3mxbzqrnhm

Prototype-based models in machine learning

Michael Biehl, Barbara Hammer, Thomas Villmann
2016 Wiley Interdisciplinary Reviews: Cognitive Science  
Supervised learning in prototype systems is exemplified in terms of learning vector quantization. Most frequently, the familiar Euclidean distance serves as a dissimilarity measure.  ...  An overview is given of prototype-based models in machine learning. In this framework, observations, i.e., data, are stored in terms of typical representatives.  ...  Metric properties are, however, not strictly required in prototype-based systems where the dissimilarity measure is used to determine the distance of a specific feature vector from the prototypes only.  ... 
doi:10.1002/wcs.1378 pmid:26800334 fatcat:tdri4kzqi5g2hkpckcbjpq7ici

Recurrent Subgraph Prediction

Saurabh Nagrecha, Nitesh V. Chawla, Horst Bunke
2015 Proceedings of the 2015 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining 2015 - ASONAM '15  
PReSub predicts re-occurring subgraphs using the network's vector space embedding and a set of "early warning subgraphs" which act as global and local descriptors of the subgraph's behavior.  ...  The evolution of these subgraphs cannot be completely predicted using a pairwise link prediction analysis.  ...  Graph Edit Distance and Prototypes In order to obtain a reduced dimensional picture of the graph, we use a graph kernel to implement vector space embedding [4] .  ... 
doi:10.1145/2808797.2809283 dblp:conf/asunam/NagrechaCB15 fatcat:mzu3haqec5ahzo6d6rczyldiv4

One-Class Classifiers Based on Entropic Spanning Graphs

Lorenzo Livi, Cesare Alippi
2017 IEEE Transactions on Neural Networks and Learning Systems  
The fuzzification process is based only on topological information of the vertices of the entropic spanning graph.  ...  The spanning graph is learned on the embedded input data and the outcoming partition of vertices defines the classifier.  ...  Embedding vectors are constructed by using a (parametric) dissimilarity measure.  ... 
doi:10.1109/tnnls.2016.2608983 pmid:28114079 fatcat:eprjle7rpzdgzhosow2xfm6wxq
« Previous Showing results 1 — 15 out of 2,018 results