Page 107 |
Save page Remove page | Previous | 107 of 196 | Next |
|
small (250x250 max)
medium (500x500 max)
Large (1000x1000 max)
Extra Large
large ( > 500x500)
Full Resolution
All (PDF)
|
This page
All
|
collection of n-sample partitions rules = { 1, 2, ...}. Let be an arbitrary parti-tion
scheme for Rd, then for every partition rule n 2 we can define its associated
collection of measurable partitions by [46]
An =
n(x1, .., xn) : (x1, .., xn) 2 Rd·n
. (4.5)
In this context, for a given n-sample partition rule n and a sequence (x1, .., xn) 2
Rd·n, n(x|x1, .., xn) denotes the mapping from any point x in Rd to its unique cell in
n(x1, .., xn), such that x 2 n(x|x1, .., xn).
4.2.3 Vapnik-Chervonenkis type of inequality
Let X1,X2, ..,Xn be independent identically distributed (i.i.d.) realizations of a random
vector with values in Rd, with X P and P a probability measure on (Rd, B(Rd)).
Then 8A 2 n(X1,X2, ..,Xn), we can define the empirical distribution by
Pn(A) =
1
n
Xn
i=1
IA(Xi), (4.6)
a probability measure defined on (Rd, ( n(X1, ..,Xn)))2. This is the abstract represen-tation
of the data-dependent partition scheme for probability estimation, where the i.i.d.
samples are used twice: first for defining a sub-sigma field ( n(X1, ..,Xn)) B(Rd)
where we want to focus the estimation problem and then for characterizing the empirical
probability measure on it.
In this learning setting, Lugosi and Nobel [46] proved a distribution free inequality to
bound the uniform deviation of the empirical distribution with respect to the probability
2 ( ) denotes the smallest sigma-field that contain , which for the case of partitions is the collection
of sets that can be written as union of cells of .
94
Object Description
| Title | On optimal signal representation for statistical learning and pattern recognition |
| Author | Silva, Jorge |
| Author email | jorgesil@usc.edu; josilva@ing.uchile.cl |
| Degree | Doctor of Philosophy |
| Document type | Dissertation |
| Degree program | Electrical Engineering |
| School | Viterbi School of Engineering |
| Date defended/completed | 2008-06-23 |
| Date submitted | 2008 |
| Restricted until | Unrestricted |
| Date published | 2008-10-21 |
| Advisor (committee chair) | Narayanan, Shrikanth S. |
| Advisor (committee member) |
Kuo, C.-C. Jay Ordóñez, Fernando I. |
| Abstract | This work presents contributions on two important aspects of the role of signal representation in statistical learning problems, in the context of deriving new methodologies and representations for speech recognition and the estimation of information theoretic quantities.; The first topic focuses on the problem of optimal filter bank selection using Wavelet Packets (WPs) for speech recognition applications. We propose new results to show an estimation-approximation error tradeoff across sequence of embedded representations. These results were used to formulate the minimum probability of error signal representation (MPE-SR) problem as a complexity regularization criterion. Restricting this criterion to the filter bank selection, algorithmic solutions are provided by exploring the dyadic tree-structure of WPs. These solutions are stipulated in terms of a set of conditional independent assumptions for the acoustic observation process, in particular, a Markov tree property across the indexed structure of WPs. In the technical side, this work presents contributions on the extension of minimum cost tree pruning algorithms and their properties to affine tree functionals. For the experimental validation, a phone classification task ratifies the goodness of Wavelet Packets as an analysis scheme for non-stationary time-series processes, and the effectiveness of the MPE-SR to provide cost effective discriminative filter bank solution for pattern recognition.; The second topic addresses the problem of data-dependent partitions for the estimation of mutual information and Kullback-Leibler divergence (KLD). This work proposes general histogram-based estimates considering non-product data-driven partition schemes. The main contribution is the stipulation of sufficient conditions to make these histogram-based constructions strongly consistent for both problems. The sufficient conditions consider combinatorial complexity indicator for partition families and the use of large deviation type of inequalities (Vapnik-Chervonenkis inequalities). On the application side, two emblematic data-dependent constructions are derived from this result, one based on statistically equivalent blocks and the other, on a tree-structured vector quantization scheme. A range of design values was stipulated to guarantee strongly consistency estimates for both framework. Furthermore, experimental results under controlled settings demonstrate the superiority of these data-driven techniques in terms of a bias-variance analysis when compared to conventional product histogram-based and kernel plug-in estimates. |
| Keyword | signal representation in statistical learning; Bayes decision theory; basis selection; tree-structured bases and Wavelet packet (WP); complexity regularization; minimum cost tree pruning; family pruning problem; mutual information estimation; divergence estimation; data-dependent partitions; statistical learning theory; concentration inequalities; tree-structured vector quantization. |
| Language | English |
| Part of collection | University of Southern California dissertations and theses |
| Publisher (of the original version) | University of Southern California |
| Place of publication (of the original version) | Los Angeles, California |
| Publisher (of the digital version) | University of Southern California. Libraries |
| Provenance | Electronically uploaded by the author |
| Type | texts |
| Legacy record ID | usctheses-m1684 |
| Contributing entity | University of Southern California |
| Rights | Silva, Jorge |
| Repository name | Libraries, University of Southern California |
| Repository address | Los Angeles, California |
| Repository email | cisadmin@lib.usc.edu |
| Filename | etd-Silva-2450 |
| Archival file | uscthesesreloadpub_Volume32/etd-Silva-2450.pdf |
Description
| Title | Page 107 |
| Contributing entity | University of Southern California |
| Repository email | cisadmin@lib.usc.edu |
| Full text | collection of n-sample partitions rules = { 1, 2, ...}. Let be an arbitrary parti-tion scheme for Rd, then for every partition rule n 2 we can define its associated collection of measurable partitions by [46] An = n(x1, .., xn) : (x1, .., xn) 2 Rd·n . (4.5) In this context, for a given n-sample partition rule n and a sequence (x1, .., xn) 2 Rd·n, n(x|x1, .., xn) denotes the mapping from any point x in Rd to its unique cell in n(x1, .., xn), such that x 2 n(x|x1, .., xn). 4.2.3 Vapnik-Chervonenkis type of inequality Let X1,X2, ..,Xn be independent identically distributed (i.i.d.) realizations of a random vector with values in Rd, with X P and P a probability measure on (Rd, B(Rd)). Then 8A 2 n(X1,X2, ..,Xn), we can define the empirical distribution by Pn(A) = 1 n Xn i=1 IA(Xi), (4.6) a probability measure defined on (Rd, ( n(X1, ..,Xn)))2. This is the abstract represen-tation of the data-dependent partition scheme for probability estimation, where the i.i.d. samples are used twice: first for defining a sub-sigma field ( n(X1, ..,Xn)) B(Rd) where we want to focus the estimation problem and then for characterizing the empirical probability measure on it. In this learning setting, Lugosi and Nobel [46] proved a distribution free inequality to bound the uniform deviation of the empirical distribution with respect to the probability 2 ( ) denotes the smallest sigma-field that contain , which for the case of partitions is the collection of sets that can be written as union of cells of . 94 |
