Eguchi shinto (26 Ergebnisse)
- Softcover
Anbieter: Books From California, Simi Valley, CA, USABooks From California
Verkäufer/-in kontaktierenVerkäufer/-in mit 4 SternenZustand: Gebraucht - Gut
EUR 29,33
EUR 4,35 VersandVersand innerhalb von USAAnzahl: 1 verfügbar
paperback. Zustand: Very Good. Cover and edges may have some wear.
- Softcover
Anbieter: Grand Eagle Retail, Bensenville, IL, USAGrand Eagle Retail
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 61,36
Versand nach gratisVersand innerhalb von USAAnzahl: 1 verfügbar
Paperback. Zustand: new. Paperback. This book introduces the gamma-divergence, a measure of distance between probability distributions that was proposed by Fujisawa and Eguchi in 2008. The gamma-divergence has been extensively explored to provide robust estimation when the power index g is positive. The gamma-divergence can be d…efined even when the power index g is negative, as long as the condition of integrability is satisfied. Thus, the authors consider the gamma-divergence defined on a set of discrete distributions. The arithmetic, geometric, and harmonic means for the distribution ratios are closely connected with the gamma-divergence with a negative g. In particular, the authors call the geometric-mean (GM) divergence the gamma-divergence when g is equal to -1.The book begins by providing an overview of the gamma-divergence and its properties. It then goes on to discuss the applications of the gamma-divergence in various areas, including machine learning, statistics, and ecology. Bernoulli, categorical, Poisson, negative binomial, and Boltzmann distributions are discussed as typical examples. Furthermore, regression analysis models that explicitly or implicitly assume these distributions as the dependent variable in generalized linear models are discussed to apply the minimum gamma-divergence method.In ensemble learning, AdaBoost is derived by the exponential loss function in the weighted majority vote manner. It is pointed out that the exponential loss function is deeply connected to the GM divergence. In the Boltzmann machine, the maximum likelihood has to use approximation methods such as mean field approximation because of the intractable computation of the partition function. However, by considering the GM divergence and the exponential loss, it is shown that the calculation of the partition function is not necessary, and it can be executed without variational inference. In particular, the authors call the geometric-mean (GM) divergence the gamma-divergence when g is equal to -1.The book begins by providing an overview of the gamma-divergence and its properties. Shipping may be from multiple locations in the US or from the UK, depending on stock availability.
- Softcover
Anbieter: Ria Christie Collections, Uxbridge, Vereinigtes KönigreichRia Christie Collections
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 53,77
EUR 14,14 VersandVersand von Vereinigtes Königreich nach USAAnzahl: Mehr als 20 verfügbar
Zustand: New. In.
- Softcover
Anbieter: Books Puddle, New York, NY, USABooks Puddle
Verkäufer/-in kontaktierenVerkäufer/-in mit 4 SternenZustand: Neu
EUR 77,21
EUR 3,48 VersandVersand innerhalb von USAAnzahl: 4 verfügbar
Zustand: New.
- Softcover
Anbieter: Revaluation Books, Exeter, Vereinigtes KönigreichRevaluation Books
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 76,30
EUR 11,80 VersandVersand von Vereinigtes Königreich nach USAAnzahl: 2 verfügbar
Paperback. Zustand: Brand New. 1st edition. 59 pages. 9.25x6.25x0.50 inches. In Stock.
- Softcover
Anbieter: AHA-BUCH GmbH, Einbeck, DeutschlandAHA-BUCH GmbH
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 58,39
EUR 61,04 VersandVersand von Deutschland nach USAAnzahl: 1 verfügbar
Taschenbuch. Zustand: Neu. Druck auf Anfrage Neuware - Printed after ordering - This book introduces the gamma-divergence, a measure of distance between probability distributions that was proposed by Fujisawa and Eguchi in 2008. The gamma-divergence has been extensively explored to provide robust estimation when the power index…Gamma is positive. The gamma-divergence can be defined even when the power index Gamma is negative, as long as the condition of integrability is satisfied. Thus, the authors consider the gamma-divergence defined on a set of discrete distributions. The arithmetic, geometric, and harmonic means for the distribution ratios are closely connected with the gamma-divergence with a negative Gamma. In particular, the authors call the geometric-mean (GM) divergence the gamma-divergence when Gamma is equal to -1.The book begins by providing an overview of the gamma-divergence and its properties. It then goes on to discuss the applications of the gamma-divergence in various areas, including machine learning, statistics, and ecology. Bernoulli, categorical, Poisson, negative binomial, and Boltzmann distributions are discussed as typical examples. Furthermore, regression analysis models that explicitly or implicitly assume these distributions as the dependent variable in generalized linear models are discussed to apply the minimum gamma-divergence method.In ensemble learning, AdaBoost is derived by the exponential loss function in the weighted majority vote manner. It is pointed out that the exponential loss function is deeply connected to the GM divergence. In the Boltzmann machine, the maximum likelihood has to use approximation methods such as mean field approximation because of the intractable computation of the partition function. However, by considering the GM divergence and the exponential loss, it is shown that the calculation of the partition function is not necessary, and it can be executed without variational inference.
- Softcover
Anbieter: AussieBookSeller, Truganina, VIC, AustralienAussieBookSeller
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 98,24
EUR 32,26 VersandVersand von Australien nach USAAnzahl: 1 verfügbar
Paperback. Zustand: new. Paperback. This book introduces the gamma-divergence, a measure of distance between probability distributions that was proposed by Fujisawa and Eguchi in 2008. The gamma-divergence has been extensively explored to provide robust estimation when the power index g is positive. The gamma-divergence can be d…efined even when the power index g is negative, as long as the condition of integrability is satisfied. Thus, the authors consider the gamma-divergence defined on a set of discrete distributions. The arithmetic, geometric, and harmonic means for the distribution ratios are closely connected with the gamma-divergence with a negative g. In particular, the authors call the geometric-mean (GM) divergence the gamma-divergence when g is equal to -1.The book begins by providing an overview of the gamma-divergence and its properties. It then goes on to discuss the applications of the gamma-divergence in various areas, including machine learning, statistics, and ecology. Bernoulli, categorical, Poisson, negative binomial, and Boltzmann distributions are discussed as typical examples. Furthermore, regression analysis models that explicitly or implicitly assume these distributions as the dependent variable in generalized linear models are discussed to apply the minimum gamma-divergence method.In ensemble learning, AdaBoost is derived by the exponential loss function in the weighted majority vote manner. It is pointed out that the exponential loss function is deeply connected to the GM divergence. In the Boltzmann machine, the maximum likelihood has to use approximation methods such as mean field approximation because of the intractable computation of the partition function. However, by considering the GM divergence and the exponential loss, it is shown that the calculation of the partition function is not necessary, and it can be executed without variational inference. In particular, the authors call the geometric-mean (GM) divergence the gamma-divergence when g is equal to -1.The book begins by providing an overview of the gamma-divergence and its properties. Shipping may be from our Sydney, NSW warehouse or from our UK or US warehouse, depending on stock availability.
- Softcover
Anbieter: Mispah books, Redhill, SURRE, Vereinigtes KönigreichMispah books
Verkäufer/-in kontaktierenVerkäufer/-in mit 4 SternenZustand: Neu
EUR 111,86
EUR 29,51 VersandVersand von Vereinigtes Königreich nach USAAnzahl: 1 verfügbar
Paperback. Zustand: New. New. book.
- Hardcover
Anbieter: Ria Christie Collections, Uxbridge, Vereinigtes KönigreichRia Christie Collections
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 142,69
EUR 14,14 VersandVersand von Vereinigtes Königreich nach USAAnzahl: Mehr als 20 verfügbar
Zustand: New. In.
Minimum Divergence Methods in Statistical Machine Learning : From an Information Geometric Viewpoint
- Hardcover
Anbieter: Buchpark, Trebbin, DeutschlandBuchpark
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Gebraucht
EUR 84,65
EUR 105,00 VersandVersand von Deutschland nach USAAnzahl: 2 verfügbar
Zustand: Hervorragend. Zustand: Hervorragend | Sprache: Englisch | Produktart: Bücher | This book explores minimum divergence methods of statistical machine learning for estimation, regression, prediction, and so forth, in which we engage in information geometry to elucidate their intrinsic properties of the corresponding loss f…unctions, learning algorithms, and statistical models. One of the most elementary examples is Gauss's least squares estimator in a linear regression model, in which the estimator is given by minimization of the sum of squares between a response vector and a vector of the linear subspace hulled by explanatory vectors. This is extended to Fisher's maximum likelihood estimator (MLE) for an exponential model, in which the estimator is provided by minimization of the Kullback-Leibler (KL) divergence between a data distribution and a parametric distribution of the exponential model in an empirical analogue. Thus, we envisage a geometric interpretation of such minimization procedures such that a right triangle is kept with Pythagorean identity in the sense of the KL divergence. This understanding sublimates a dualistic interplay between a statistical estimation and model, which requires dual geodesic paths, called m-geodesic and e-geodesic paths, in a framework of information geometry. We extend such a dualistic structure of the MLE and exponential model to that of the minimum divergence estimator and the maximum entropy model, which is applied to robust statistics, maximum entropy, density estimation, principal component analysis, independent component analysis, regression analysis, manifold learning, boosting algorithm, clustering, dynamic treatment regimes, and so forth. We consider a variety of information divergence measures typically including KL divergence to express departure from one probability distribution to another. An information divergence is decomposed into the cross-entropy and the (diagonal) entropy in which the entropy associates with a generative model as a family of maximum entropy distributions; the cross entropy associates with a statistical estimation method via minimization of the empirical analogue based on given data. Thus any statistical divergence includes an intrinsic object between the generative model and the estimation method. Typically, KL divergence leads to the exponential model and the maximum likelihood estimation. It is shown that any information divergence leads to a Riemannian metric and a pair of the linear connections in the framework of information geometry. We focus on a class of information divergence generated by an increasing and convex function U, called U-divergence. It is shown that any generator function U generates the U-entropy and U-divergence, in which there is a dualistic structure between the U-divergence method and the maximum U-entropy model. We observe that a specific choice of U leads to a robust statistical procedurevia the minimum U-divergence method. If U is selected as an exponential function, then the corresponding U-entropy and U-divergence are reduced to the Boltzmann-Shanon entropy and the KL divergence; the minimum U-divergence estimator is equivalent to the MLE. For robust supervised learning to predict a class label we observe that the U-boosting algorithm performs well for contamination of mislabel examples if U is appropriately selected. We present such maximal U-entropy and minimum U-divergence methods, in particular, selecting a power function as U to provide flexible performance in statistical machine l.
- Hardcover
Anbieter: Books Puddle, New York, NY, USABooks Puddle
Verkäufer/-in kontaktierenVerkäufer/-in mit 4 SternenZustand: Neu
EUR 215,46
EUR 3,48 VersandVersand innerhalb von USAAnzahl: 4 verfügbar
Zustand: New. 1st ed. 2022 edition NO-PA16APR2015-KAP.
- Hardcover
Anbieter: BUCHSERVICE / ANTIQUARIAT Lars Lutzer, Wahlstedt, DeutschlandBUCHSERVICE / ANTIQUARIAT Lars Lutzer
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Gebraucht - Gut
EUR 189,00
EUR 39,95 VersandVersand von Deutschland nach USAAnzahl: 1 verfügbar
Hardcover. Zustand: gut. 2022. Minimum Divergence Methods in Statistical Machine Learning In deutscher Sprache. pages.
Minimum Divergence Methods in Statistical Machine Learning : From an Information Geometric Viewpoint
- Hardcover
Anbieter: AHA-BUCH GmbH, Einbeck, DeutschlandAHA-BUCH GmbH
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 168,73
EUR 62,59 VersandVersand von Deutschland nach USAAnzahl: 1 verfügbar
Buch. Zustand: Neu. Druck auf Anfrage Neuware - Printed after ordering - This book explores minimum divergence methods of statistical machine learning for estimation, regression, prediction, and so forth, in which we engage in information geometry to elucidate their intrinsic properties of the corresponding loss functions, learn…ing algorithms, and statistical models.One of the most elementary examples is Gauss's least squares estimator in a linear regression model, in which the estimator is given by minimization of the sum of squares between a response vector and a vector of the linear subspace hulled by explanatory vectors. This is extended to Fisher's maximum likelihood estimator (MLE) for an exponential model, in which the estimator is provided by minimization of the Kullback-Leibler (KL) divergence between a data distribution and a parametric distribution of the exponential model in an empirical analogue. Thus, we envisage a geometric interpretation of such minimization procedures such that a right triangle is kept with Pythagorean identity in the sense of the KL divergence. This understanding sublimates a dualistic interplay between a statistical estimation and model, which requires dual geodesic paths, called m-geodesic and e-geodesic paths, in a framework of information geometry. We extend such a dualistic structure of the MLE and exponential model to that of the minimum divergence estimator and the maximum entropy model, which is applied to robust statistics, maximum entropy, density estimation, principal component analysis, independent component analysis, regression analysis, manifold learning, boosting algorithm, clustering, dynamic treatment regimes, and so forth. We consider a variety of information divergence measures typically including KL divergence to express departure from one probability distribution to another. An information divergence is decomposed into the cross-entropy and the (diagonal) entropy in which the entropy associates with a generative model as a family of maximum entropy distributions; the cross entropy associates with a statistical estimation method via minimization of the empirical analogue based on given data. Thus any statistical divergence includes an intrinsic object between the generative model and the estimation method. Typically, KL divergence leads to the exponential model and the maximum likelihood estimation. It is shown that any information divergence leads to a Riemannian metric and a pair of the linear connections in the framework of information geometry. We focus on a class of information divergence generated by an increasing and convex function U, called U-divergence. It is shown that any generator function U generates the U-entropy and U-divergence, in which there is a dualistic structure between the U-divergence method and the maximum U-entropy model. We observe that a specific choice of U leads to a robust statistical procedurevia the minimum U-divergence method. If U is selected as an exponential function, then the corresponding U-entropy and U-divergence are reduced to the Boltzmann-Shanon entropy and the KL divergence; the minimum U-divergence estimator is equivalent to the MLE. For robust supervised learning to predict a class label we observe that the U-boosting algorithm performs well for contamination of mislabel examples if U is appropriately selected. We present such maximal U-entropy and minimum U-divergence methods, in particular, selecting a power function as U to provide flexible performance in statistical machine learning.
- Softcover
- Print-on-Demand
Anbieter: Revaluation Books, Exeter, Vereinigtes KönigreichRevaluation Books
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 48,62
EUR 11,80 VersandVersand von Vereinigtes Königreich nach USAAnzahl: 1 verfügbar
Paperback. Zustand: Brand New. 126 pages. 9.25x6.10x9.21 inches. In Stock. This item is printed on demand.
Sprache: Englisch
Verlag: Springer, Berlin, Springer Nature Singapore, Springer, 2025
- Softcover
- Print-on-Demand
Anbieter: BuchWeltWeit Ludwig Meier e.K., Bergisch Gladbach, DeutschlandBuchWeltWeit Ludwig Meier e.K.
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 48,14
EUR 23,00 VersandVersand von Deutschland nach USAAnzahl: 2 verfügbar
Taschenbuch. Zustand: Neu. This item is printed on demand - it takes 3-4 days longer - Neuware -This book introduces the gamma-divergence, a measure of distance between probability distributions that was proposed by Fujisawa and Eguchi in 2008. The gamma-divergence has been extensively explored to provide robust estimation when…the power index Gamma is positive. The gamma-divergence can be defined even when the power index Gamma is negative, as long as the condition of integrability is satisfied. Thus, the authors consider the gamma-divergence defined on a set of discrete distributions. The arithmetic, geometric, and harmonic means for the distribution ratios are closely connected with the gamma-divergence with a negative Gamma. In particular, the authors call the geometric-mean (GM) divergence the gamma-divergence when Gamma is equal to -1.The book begins by providing an overview of the gamma-divergence and its properties. It then goes on to discuss the applications of the gamma-divergence in various areas, including machine learning, statistics, and ecology. Bernoulli, categorical, Poisson, negative binomial, and Boltzmann distributions are discussed as typical examples. Furthermore, regression analysis models that explicitly or implicitly assume these distributions as the dependent variable in generalized linear models are discussed to apply the minimum gamma-divergence method.In ensemble learning, AdaBoost is derived by the exponential loss function in the weighted majority vote manner. It is pointed out that the exponential loss function is deeply connected to the GM divergence. In the Boltzmann machine, the maximum likelihood has to use approximation methods such as mean field approximation because of the intractable computation of the partition function. However, by considering the GM divergence and the exponential loss, it is shown that the calculation of the partition function is not necessary, and it can be executed without variational inference. 118 pp. Englisch.
- Softcover
- Print-on-Demand
Anbieter: BuchWeltWeit Ludwig Meier e.K., Bergisch Gladbach, DeutschlandBuchWeltWeit Ludwig Meier e.K.
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 53,49
EUR 23,00 VersandVersand von Deutschland nach USAAnzahl: 2 verfügbar
Taschenbuch. Zustand: Neu. This item is printed on demand - it takes 3-4 days longer - Neuware -This book presents a fresh, new approach in that it provides a comprehensive recent review of challenging problems caused by imbalanced data in prediction and classification, and also in that it introduces several of the latest statis…tical methods of dealing with these problems. The book discusses the property of the imbalance of data from two points of view. The first is quantitative imbalance, meaning that the sample size in one population highly outnumbers that in another population. It includes presence-only data as an extreme case, where the presence of a species is confirmed, whereas the information on its absence is uncertain, which is especially common in ecology in predicting habitat distribution. The second is qualitative imbalance, meaning that the data distribution of one population can be well specified whereas that of the other one shows a highly heterogeneous property. A typical case is the existence of outliers commonly observed in gene expression data, and another is heterogeneous characteristics often observed in a case group in case-control studies. The extension of the logistic regression model, maxent, and AdaBoost for imbalanced data is discussed, providing a new framework for improvement of prediction, classification, and performance of variable selection. Weights functions introduced in the methods play an important role in alleviating the imbalance of data. This book also furnishes a new perspective on these problem and shows some applications of the recently developed statistical methods to real data sets. 68 pp. Englisch.
- Softcover
- Print-on-Demand
Anbieter: Majestic Books, Hounslow, Vereinigtes KönigreichMajestic Books
Verkäufer/-in kontaktierenVerkäufer/-in mit 4 SternenZustand: Neu
EUR 78,13
EUR 7,67 VersandVersand von Vereinigtes Königreich nach USAAnzahl: 4 verfügbar
Zustand: New. Print on Demand.
- Softcover
- Print-on-Demand
Anbieter: Biblios, frankfurt am main, HESSE, DeutschlandBiblios
Verkäufer/-in kontaktierenVerkäufer/-in mit 4 SternenZustand: Neu
EUR 78,87
EUR 9,95 VersandVersand von Deutschland nach USAAnzahl: 4 verfügbar
Zustand: New. PRINT ON DEMAND.
- Softcover
- Print-on-Demand
Anbieter: moluna, Greven, Deutschlandmoluna
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 48,37
EUR 48,99 VersandVersand von Deutschland nach USAAnzahl: Mehr als 20 verfügbar
Zustand: New. Dieser Artikel ist ein Print on Demand Artikel und wird nach Ihrer Bestellung fuer Sie gedruckt.
- Softcover
- Print-on-Demand
Anbieter: moluna, Greven, Deutschlandmoluna
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 48,37
EUR 48,99 VersandVersand von Deutschland nach USAAnzahl: Mehr als 20 verfügbar
Zustand: New. Dieser Artikel ist ein Print on Demand Artikel und wird nach Ihrer Bestellung fuer Sie gedruckt. Osamu Komori, The Institute of Statistical Mathematics, Shinto Eguchi, The Institute of Statistical MathematicsThis book presents a fresh, new approach in that it provides a comprehensive recent review of challengin…g problems caused by imba.
- Softcover
- Print-on-Demand
Anbieter: buchversandmimpf2000, Emtmannsberg, BAYE, Deutschlandbuchversandmimpf2000
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 53,49
EUR 60,00 VersandVersand von Deutschland nach USAAnzahl: 1 verfügbar
Taschenbuch. Zustand: Neu. This item is printed on demand - Print on Demand Titel. Neuware -This book introduces the gamma-divergence, a measure of distance between probability distributions that was proposed by Fujisawa and Eguchi in 2008. The gamma-divergence has been extensively explored to provide robust estimation when the…power index ¿ is positive. The gamma-divergence can be defined even when the power index ¿ is negative, as long as the condition of integrability is satisfied. Thus, the authors consider the gamma-divergence defined on a set of discrete distributions. The arithmetic, geometric, and harmonic means for the distribution ratios are closely connected with the gamma-divergence with a negative ¿. In particular, the authors call the geometric-mean (GM) divergence the gamma-divergence when ¿ is equal to -1.The book begins by providing an overview of the gamma-divergence and its properties. It then goes on to discuss the applications of the gamma-divergence in various areas, including machine learning, statistics, and ecology. Bernoulli, categorical, Poisson, negative binomial, and Boltzmann distributions are discussed as typical examples. Furthermore, regression analysis models that explicitly or implicitly assume these distributions as the dependent variable in generalized linear models are discussed to apply the minimum gamma-divergence method.In ensemble learning, AdaBoost is derived by the exponential loss function in the weighted majority vote manner. It is pointed out that the exponential loss function is deeply connected to the GM divergence. In the Boltzmann machine, the maximum likelihood has to use approximation methods such as mean field approximation because of the intractable computation of the partition function. However, by considering the GM divergence and the exponential loss, it is shown that the calculation of the partition function is not necessary, and it can be executed without variational inference.Springer-Verlag KG, Sachsenplatz 4-6, 1201 Wien 128 pp. Englisch.
- Hardcover
- Print-on-Demand
Anbieter: Brook Bookstore On Demand, Napoli, NA, ItalienBrook Bookstore On Demand
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 126,26
EUR 6,80 VersandVersand von Italien nach USAAnzahl: Mehr als 20 verfügbar
Zustand: new. Questo è un articolo print on demand.
- Hardcover
- Print-on-Demand
Anbieter: moluna, Greven, Deutschlandmoluna
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 136,16
EUR 48,99 VersandVersand von Deutschland nach USAAnzahl: Mehr als 20 verfügbar
Zustand: New. Dieser Artikel ist ein Print on Demand Artikel und wird nach Ihrer Bestellung fuer Sie gedruckt. Provides various applications including boosting and kernel methods in machine learning with a geometric invariance viewpointFacilitates a deeper understanding of the complementary relation between statistical models an…d estimation in the context .
- Hardcover
- Print-on-Demand
Anbieter: buchversandmimpf2000, Emtmannsberg, BAYE, Deutschlandbuchversandmimpf2000
Verkäufer/-in kontaktierenVerkäufer/-in mit 5 SternenZustand: Neu
EUR 160,49
EUR 60,00 VersandVersand von Deutschland nach USAAnzahl: 1 verfügbar
Buch. Zustand: Neu. This item is printed on demand - Print on Demand Titel. Neuware -This book explores minimum divergence methods of statistical machine learning for estimation, regression, prediction, and so forth, in which we engage in information geometry to elucidate their intrinsic properties of the corresponding loss func…tions, learning algorithms, and statistical models. One of the most elementary examples is Gauss's least squares estimator in a linear regression model, in which the estimator is given by minimization of the sum of squares between a response vector and a vector of the linear subspace hulled by explanatory vectors. This is extended to Fisher's maximum likelihood estimator (MLE) for an exponential model, in which the estimator is provided by minimization of the Kullback-Leibler (KL) divergence between a data distribution and a parametric distribution of the exponential model in an empirical analogue. Thus, we envisage a geometric interpretation of such minimization procedures such that a right triangle is kept with Pythagorean identity in the sense of the KL divergence. This understanding sublimates a dualistic interplay between a statistical estimation and model, which requires dual geodesic paths, called m-geodesic and e-geodesic paths, in a framework of information geometry.We extend such a dualistic structure of the MLE and exponential model to that of the minimum divergence estimator and the maximum entropy model, which is applied to robust statistics, maximum entropy, density estimation, principal component analysis, independent component analysis, regression analysis, manifold learning, boosting algorithm, clustering, dynamic treatment regimes, and so forth. We consider a variety of information divergence measures typically including KL divergence to express departure from one probability distribution to another. An information divergence is decomposed into the cross-entropy and the (diagonal) entropy in which the entropy associates with a generative model as a family of maximum entropy distributions; the cross entropy associates with a statistical estimation method via minimization of the empirical analogue based on given data. Thus any statistical divergence includes an intrinsic object between the generative model and the estimation method. Typically, KL divergence leads to the exponential model and the maximum likelihood estimation. It is shown that any information divergence leads to a Riemannian metric and a pair of the linear connections in the framework of information geometry.We focus on a class of information divergence generated by an increasing and convex function U, called U-divergence. It is shown that any generator function U generates the U-entropy and U-divergence, in which there is a dualistic structure between the U-divergence method and the maximum U-entropy model. We observe that a specific choice of U leads to a robust statistical procedurevia the minimum U-divergence method. If U is selected as an exponential function, then the corresponding U-entropy and U-divergence are reduced to the Boltzmann-Shanon entropy and the KL divergence; the minimum U-divergence estimator is equivalent to the MLE. For robust supervised learning to predict a class label we observe that the U-boosting algorithm performs well for contamination of mislabel examples if U is appropriately selected. We present such maximal U-entropy and minimum U-divergence methods, in particular, selecting a power function as U to provide flexible performance in statistical machine learning.Springer-Verlag KG, Sachsenplatz 4-6, 1201 Wien 232 pp. Englisch.
- Hardcover
- Print-on-Demand
Anbieter: Majestic Books, Hounslow, Vereinigtes KönigreichMajestic Books
Verkäufer/-in kontaktierenVerkäufer/-in mit 4 SternenZustand: Neu
EUR 228,16
EUR 7,67 VersandVersand von Vereinigtes Königreich nach USAAnzahl: 4 verfügbar
Zustand: New. Print on Demand.
- Hardcover
- Print-on-Demand
Anbieter: Biblios, frankfurt am main, HESSE, DeutschlandBiblios
Verkäufer/-in kontaktierenVerkäufer/-in mit 4 SternenZustand: Neu
EUR 229,12
EUR 9,95 VersandVersand von Deutschland nach USAAnzahl: 4 verfügbar
Zustand: New. PRINT ON DEMAND.
