Berg, C., J. P. R. Christensen, and P. Ressel, Harmonic Analysis on Semigroups, Springer-Verlag, New York, 1984.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....# ij , d ij , # ik , d ik , # jk and d jk , # ik =# d ij d jk d ik . Proof of Lemma 1. d ij d jk ) # ij # jk 2# ij # jk # ik = d ik . Hence, d ij d jk d ik . # Lemma 2 The matrix A, whose (i, j) th element is a ij = is Euclidean. Proof of Lemma 2. By [18], E A is positive semi definite, where E is a matrix with all elements being 1. By the result from [10] A is Euclidean. # other words, d ij satisfies the following properties: j, k. Proof of Theorem 1. The first three properties are obvious; we will only prove the triangular ....

C. Berg, J. P. R. Christensen, and P. Ressel, Harmonic Analysis on Semigroups. New York: Springer-Verlag, 1984. 18

....function k(x, y) to represent k : X X # F , k(x, y) #(x) # #(y) which allows us to compute the value of the inner product in F without carrying out the map #. This is sometimes referred as the kernel trick . In most cases, we pay much attention to positive definite kernels. Definition 1 [16] Let be a nonempty set. A function k : X X # R is a positive definite kernel if and only if i,j=1 a i a j k(x j , x k ) 0, for all n # N a 1 , a n R. If is a finite set (X = say) then k is positive definite if and only if the n n matrix Gram matrix ....

....= k 1 (x 1 , x 2 )k 2 (u 1 , u 2 ) is called the tensor product kernel . Similarly, their direct sum, k 1 U) k 2 ( x 1 , u 1 ) x 2 , u 2 ) k 1 (x 1 , x 2 ) k 2 (u 1 , u 2 ) is called the direct sum kernel . Both kernels enjoy the important property that they are positive definite [16, 15, 6]. 3. Euclidean Embedding with Kernels Generally, a training data set consists of a set of observations and their corresponding labels or targets. Therefore, the training data set is not Euclidean. In this Section, we embed this set into an Euclidean space by using kernel operations. 3.1. Using ....

C. Berg, J. P. R. Christensen, and P. Ressel, Harmonic Analysis on Semigroups. New York: Springer-Verlag, 1984.

....above holds true, without necessarily worrying about the actual form of already the existence of the linear space F facilitates a number of algorithmic and theoretical issues. It is well established that (1) works out for Mercer kernels [3, 13] or, equivalently, positive definite kernels [2, 14]. Here and below, indices i and j by default run over 1; m. Definition 1 (Positive definite kernel) A symmetric function k : X X R which for all m 2 N; x i 2 X gives rise to a positive definite Gram matrix, i.e. for which for all c i 2 R we have c i c j K ij 0; where K ij : k(x ....

....Gram matrix, i.e. for which for all c i 2 R we have c i c j K ij 0; where K ij : k(x i ; x j ) 2) is called a positive definite (pd) kernel. One particularly intuitive way to construct a feature map satisfying (1) for such a kernel k proceeds, in a nutshell, as follows (for details, see [2]) 1. Define a feature map : X R ; x 7 k( x) 3) Here, R denotes the space of functions mapping X into R. 2. Turn it into a linear space by forming linear combinations f( i k( x i ) g( j k( x j ) m; m 2 N; i ; j 2 R; x i ; x j 2 X ) 4) 3. ....

[Article contains additional citation context not shown here]

C. Berg, J.P.R. Christensen, and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, New York, 1984.

....the classical scaling algorithm on the transformed dissimilarities. Critchley suggests the power transformation f( ij ) ij (for 0) If the dissimilarities are derived from Euclidean distances, we note that the kernel k(x; y) jjx yjj is conditionally positive de nite (CPD) if 2 [1]. When the kernel is CPD, the centered matrix will be positive de nite. Critchley s use of the classical scaling algorithm is similar to the algorithm discussed below, but crucially the kernel PCA method ensures that the matrix B derived form the transformed dissimilarities is non negative de ....

C. Berg, J. P. R. Christensen, and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, New York, 1984.

....the dot product in that feature space. This has been brought to the attention of the machine learning community by [1] 9] and [39] In functional analysis, the issue has been studied under the heading of Hilbert space representations of kernels. A good monograph on the theory of kernels is [5]. Besides (52) 9] and [39] suggest the usage of Gaussian radial basis function kernels [1] 58) and sigmoid kernels tanh 3 (59) where , and are real parameters. The examples given so far apply to the case of vectorial data. In ....

C. Berg, J. P. R. Christensen, and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, New York, 1984.

....feature space This question has been brought to the attention of the machine learning community by [1, 8, 28] In functional analysis, the same problem has been studied under the heading of Hilbert space representations of kernels. A good monograph on the functional analytic theory of kernels is [5]; indeed, a large part of the material in the present section is based on that work. There is one more aspect in which this section differs from the previous one: the latter dealt with vectorial data. The results in the current section, in contrast, hold for data drawn from domains which need no ....

....of Phi, which follows directly from the definition: for all functions (80) we have hk( x) fi = f(x) 86) k is the representer of evaluation. In particular, hk( x) k( x )i = k(x; x ) 87) By virtue of these properties, positive kernels k are also called reproducing kernels [3, 5, 33, 17]. By (86) and Proposition 5, we have jf(x)j = jhk( x) f ij k(x; x) Delta hf; fi: 88) Therefore, hf; fi = 0 directly implies f = 0, which is the last property that was left to prove in order to establish that h: i is a dot product. One can complete the space of functions (80) in ....

C. Berg, J.P.R. Christensen, and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, New York, 1984.

....n 1 not every nonnegative polynomial in R n can be written as a sum of squares of polynomials (see, for instance, 2, 6.3] the moment problems in n variables are more di#cult than the classical one variable problems. This very intriguing territory has been investigated by many authors (see [2], 7] 12] and their references) although characterizations for measures whose support lies in an arbitrary (generally unbounded) semi algebraic set do not seem to exist. The present paper starts from an idea of the second author, see [19] about solving moment problems by a change of basis via ....

....positive polynomials. This scheme can obviously be applied to other embeddings of the a#ne space, with similar consequences. 2. Returning to Theorem 2.8 and Corollary 4.4, we remark that the semigroup t # #(t) # , # # Z n , # # Z , is finitely generated. Consequently Theorem 6.1. 11 of [2] applies and it shows that any polynomial p satisfying p(t) # 0, t # R n , can be approximated in the finest locally convex topology of R[t, #] by elements in the convex cone # of squares of polynomials multiplied by powers of #. Indeed, p(t) #(1 #t# 2 ) d , with # 0 and 2d ....

C. Berg, J. P. R. Christensen, and P. Ressel, Harmonic Analysis on Semigroups, Grad. Texts in Math. 100, Springer-Verlag, New York, 1984.

....defined by fc = c 1 ; c 2 ; c n ) 2 IR n : P n i=1 c i = 0g. In particular, any positive definite function as defined by Schoenberg ( 11] is conditionally positive definite but not conversely. Conditionally positive definite functions in their full generality are studied in detail in [3]. For an interesting discussion on conditionally positive definite functions on homogeneous spaces see [7] The conditionally positive definite functions on the finite dimensional sphere S m were characterized (see [7] as those functions of the form g(t) g(0) Gamma X k2K a k (1 Gamma p ....

.... Gamma X k2K a k (1 Gamma p k (cos t) K ae IN n f0g; m Gamma 1) 2; a k 0; X k2K a k 1; in which p k ( Delta) P k ( Delta) P k (1) and P k denotes the standard Gegenbauer or ultraspherical polynomial. On S 1 , the representation above takes the form (see [10] and [3]) g(t) g(0) Gamma X k2K a k (1 Gamma cos k t) K ae IN n f0g; a k 0; X k2K a k 1: If g(0) P k2K a k then g is in fact a positive definite function as characterized by Schoenberg in [11] Notice that the restriction of a conditionally positive definite function on S m to D(S; ....

Berg, C., J. P. R. Christensen, and P. Ressel, "Harmonic Analysis on Semigroups", SpringerVerlag, New York, 1984.

....P 1 m;n=0 a m;n 1. The continuous functions g for which g ffi h Delta; Deltai is conditionally negative definite on S 1 are of the form g(i) g(1) 1 X m;n=0 a m;n i 1 Gamma i m i n j ; in which a m;n 0 and P 1 m;n=0 a m;n 1. A quick proof of this uses exercise 5 4. 10 in [2]. If g has any one of these series representations we write K(g) to denote the set of all pairs (m; n) for which the coefficient a m;n in the series is positive. We construct strictly positive definite and strictly conditionally negative definite kernels on S 1 employing completely monotone ....

Berg, C., J. P. R. Christensen, and P. Ressel, Harmonic Analysis on Semigroups, Springer-Verlag, New York, 1984. 12

....On the unit circle S 1 , let d 1 be the geodesic distance. The purpose of this paper is to address the problem of finding a continuous function f : 0; Gamma IR for which f ffi d 1 is either a strictly positive definite or a strictly conditionally negative definite kernel. Following [1], we say that a function f : S 1 Theta S 1 Gamma IR is a positive definite kernel if and only if n X i;j=1 c i c j f(x i ; x j ) 0 for all n 2 IN, fx 1 ; x 2 ; x n g ae S 1 , and fc 1 ; c 2 ; c n g ae IR. We say that the function f is a conditionally negative ....

Berg, C., J. P. R. Christensen, and P. Ressel, Harmonic Analysis on Semigroups, Springer Verlag, New York, 1984.

No context found.

Berg, C., J. P. R. Christensen, and P. Ressel, Harmonic Analysis on Semigroups, Springer-Verlag, New York, 1984.

No context found.

J. P. R. Christensen C. Berg and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, 1984.

No context found.

Berg Ch., Christensen J.P.R. and Ressel P. (1984) Harmonic Analysis on Semigroups. Springer-Verlag, New York.

No context found.

C. Berg, J. P. R. Christensen, and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, New York, 1984.

No context found.

C. Berg, J. P. R. Christensen, and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, 1984.

No context found.

C. Berg, J. P. R. Christensen, and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, 1984.

No context found.

Berg, C., J. P. R. Christensen, and P. Ressel (1984). Harmonic Analysis on Semigroups.

No context found.

C. Berg, J.P. Reus Christensen and P. Ressel (1984) Harmonic analysis on Semigroups. SpringerVerlag, New York

No context found.

C. Berg, J.P.R. Christensen, and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, 1984.

No context found.

C. Berg, J. P. R. Christensen, and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, New York, 1984.

No context found.

C. Berg, J. P. R. Christensen, and P. Ressel, Harmonic Analysis on Semigroups. New York: Springer-Verlag, 1984.

No context found.

BERG,C.,REUS CHRISTENSEN,J.P.andRESSEL, P. (1984). Harmonic Analysis on Semigroups. Springer, New York.

No context found.

C. Berg, J.P.R. Christensen, and P. Ressel, Harmonic Analysis on Semigroups. Springer-Verlag, 1984.

No context found.

C. Berg, J.P.R. Christensen, and P. Ressel. Harmonic Analysis on Semigroups. Springer-Verlag, 1984.

No context found.

C. Berg, J. P. R. Christensen, and P. Ressel, Harmonic Analysis on Semigroups, Springer-Verlag, New York, 1984.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC