Nonparametric Bayes error estimation using unclassified samples

Fukunaga, Keinosuke; Kessell, David

doi:10.1109/cdc.1972.269066

Cited by 9 publications

(13 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is a very powerful rule due to the fact that, for large enough training datasets, the error rate is upper bounded by twice the Bayes error rate (optimal error rate). It has also been shown that the gap between its error rate and the optimal Bayes error is directly linked to the value of k, assuming that a large enough training dataset is available [1], [2]. This is an important advantage compared to the continuous estimates such as the Gaussian mixtures.…”

Section: K-nearest Neighbour Classificationmentioning

confidence: 96%

Context-independent phoneme recognition using a K-Nearest Neighbour classification approach

Golipour

O’Shaughnessy

2009

2009 IEEE International Conference on Acoustics, Speech and Signal Processing

View full text Add to dashboard Cite

In this paper we investigate a non-parametric classification of English phonemes in speaker-independent continuous speech. We employ the "voting" k-Nearest Neighbour (k-NN) classifier, a powerful technique in pattern recognition problems, along with a new representation of phonemes for the speech recognition task. We also exploit the idea behind "approximate" k-NN that results in a very fast way of computing the k approximate closest neighbours of each data point. Comparing the recognition performance of the proposed method with the HMM-based recognizer of HTK toolkit reveals that the k-NN-based recognizer outperforms its counterpart. In addition, incorporating the "approximate" nearest neighbour search instead of the "exact" one results in completing the training step much faster than the HMM-based system, and the testing step with a comparable computational time. We also reduced the amount of the training data by applying a pattern recognition technique, called "thinning" algorithm. The outcome was a considerable reduction in the k-NN search space and hence the execution time, and also a slight increase in the recognition performance.

show abstract

Section: K-nearest Neighbour Classificationmentioning

confidence: 96%

Context-independent phoneme recognition using a K-Nearest Neighbour classification approach

Golipour

O’Shaughnessy

2009

2009 IEEE International Conference on Acoustics, Speech and Signal Processing

View full text Add to dashboard Cite

show abstract

“…An estimate Pi for the proport1on Pi may be obta1ned as follows. (5), (6), and (9), the following is obtained. .t-1 m.1.…”

Section: A10chlne Processing Of Remotely Sensed Data Symposiummentioning

confidence: 99%

Maximum Likelihood Estimation of Label Imperfection Probabilities and Its Use in the Identification of Mislabeled Patterns

Chittineni

1982

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

This paper presents the problem of estimating label imperfections and the use of the estimation in identifying mislabeled patterns. Expressions for the maximum likelihood estimates of classification errors and a priori probabilities are derived from the classification of a set of labeled and unlabeled patterns. Expressions also are presented for the asymptotic variances of probabi 1 i ty of correct class ifi cat i on and proportions. Simple models are developed for imperfections in the labels and for claSSification errors and are used in the formulation of a maximum likelihood estimation scheme. Schemes are presented for the identification of mislabeled patterns in terms of threSholds on the discriminant functions for both two-class and multiclass cases. Expressions are derived for the probability that the imperfect label identification scheme will result in a wrong decision and are used in computing thresholds. Furthermore, the results of practical applications of these techniques in the processing of remotely sensed multispectral data are presented.

show abstract

“…The other problem associated with the mixture model is to estimate a discriminant function from the unclassified data and study the performance of the discriminant function. Studies in this area have been under taken by Choi (1969), Fukunga and Kassell (1973), Moore et al (1976), Ganasalingam and Mclachlan (1978, 1979), and O'Neill (1978. In all these studies, the underlying populations are assumed to be normal.…”

Section: Introductionmentioning

confidence: 99%

Nonlinear Discriminant Functions for Mixed Random Walk Models

Moustafa

Ramadan

Mohammed

2010

Communications in Statistics - Simulation and Computation

View full text Add to dashboard Cite

A procedure is presented for finding maximum likelihood estimates of the parameters of a mixture of two random walk distributions in two cases, using classified and unclassified observations. Based on small sample size, estimation of nonlinear discriminant functions is considered. Throughout simulation experiments, the performance of the corresponding estimated nonlinear discriminant functions is investigated. The total probabilities of misclassification and percentage biases are evaluated and discussed.

show abstract

Nonparametric Bayes error estimation using unclassified samples

Cited by 9 publications

References 23 publications

Context-independent phoneme recognition using a K-Nearest Neighbour classification approach

Context-independent phoneme recognition using a K-Nearest Neighbour classification approach

Maximum Likelihood Estimation of Label Imperfection Probabilities and Its Use in the Identification of Mislabeled Patterns

Nonlinear Discriminant Functions for Mixed Random Walk Models

Contact Info

Product

Resources

About