On the mean accuracy of statistical pattern recognizers

Authors:
G. Hughes
Affiliations:
-
Venue:
IEEE Transactions on Information Theory
Year:
2006

Citing 0
Cited 42

Urban monitoring using multi-temporal SAR and multi-spectral data

Pattern Recognition Letters - Special issue: Pattern recognition in remote sensing (PRRS 2004)
An ensemble-driven k-NN approach to ill-posed classification problems

Pattern Recognition Letters - Special issue: Pattern recognition in remote sensing (PRRS 2004)
Partition based pattern synthesis technique with efficient algorithms for nearest neighbor classification

Pattern Recognition Letters
Determining the saliency of feature measurements obtained from images of sedimentary organic matter for use in its classification

Computers & Geosciences
Active learning for microarray data

International Journal of Approximate Reasoning
A fast separability-based feature-selection method for high-dimensional remotely sensed image classification

Pattern Recognition
Application of wrapper approach and composite classifier to the stock trend prediction

Expert Systems with Applications: An International Journal
A fast separability-based feature-selection method for high-dimensional remotely sensed image classification

Pattern Recognition
The peaking phenomenon in the presence of feature-selection

Pattern Recognition Letters
A projection pursuit algorithm for anomaly detection in hyperspectral imagery

Pattern Recognition
Performance of feature-selection methods in the classification of high-dimension data

Pattern Recognition
Exact correlation between actual and estimated errors in discrete classification

Pattern Recognition Letters
Multiband segmentation based on a hierarchical Markov model

Pattern Recognition
Implementation of real-time constrained linear discriminant analysis to remote sensing image classification

Pattern Recognition
Overlap pattern synthesis with an efficient nearest neighbor classifier

Pattern Recognition
Exact performance of error estimators for discrete classifiers

Pattern Recognition
Segmentation and classification of hyperspectral images using watershed transformation

Pattern Recognition
SmartSpectra: Applying multispectral imaging to industrial environments

Real-Time Imaging
Rasch-based high-dimensionality data reduction and class prediction with applications to microarray gene expression data

Expert Systems with Applications: An International Journal
Applications of machine learning techniques to a sensor-network-based prosthesis training system

Applied Soft Computing
Classification of healthy and abnormal swallows based on accelerometry and nasal airflow signals

Artificial Intelligence in Medicine
A spatial-spectral kernel-based approach for the classification of remote-sensing images

Pattern Recognition
Robust fitting of fluorescence spectra for pre-symptomatic wheat leaf rust detection with Support Vector Machines

Computers and Electronics in Agriculture
A supervised clustering approach for fMRI-based inference of brain states

Pattern Recognition
Two-stage phone duration modelling with feature construction and feature vector extension for the needs of speech synthesis

Computer Speech and Language
Computer networking for the robotic neural and sensory systems

Computer Communications
Segmentation and detection of colorectal polyps using local polynomial approximation

ICIAR'12 Proceedings of the 9th international conference on Image Analysis and Recognition - Volume Part II
Tongue-print: a novel biometrics pattern

ICB'07 Proceedings of the 2007 international conference on Advances in Biometrics
Cascaded classification of high resolution remote sensing images using multiple contexts

Information Sciences: an International Journal
Fast and efficient lung disease classification using hierarchical one-against-all support vector machine and cost-sensitive feature selection

Computers in Biology and Medicine
The dipping phenomenon

SSPR'12/SPR'12 Proceedings of the 2012 Joint IAPR international conference on Structural, Syntactic, and Statistical Pattern Recognition
Optimal classifiers with minimum expected error within a Bayesian framework-Part I: Discrete and Gaussian models

Pattern Recognition
Target Tracking Using Multiple Patches and Weighted Vector Median Filters

Journal of Mathematical Imaging and Vision
Survival model in oral squamous cell carcinoma based on clinicopathological parameters, molecular markers and support vector machines

Expert Systems with Applications: An International Journal
Resampling methods for quality assessment of classifier performance and optimal number of features

Signal Processing
Generating supplemental content information using virtual profiles

Proceedings of the 7th ACM conference on Recommender systems
Target detection based on a dynamic subspace

Pattern Recognition
Community detection and visualization in social networks: Integrating structural and semantic information

ACM Transactions on Intelligent Systems and Technology (TIST) - Special Section on Intelligent Mobile Knowledge Discovery and Management Systems and Special Issue on Social Web Mining
Multiobjective genetic programming for maximizing ROC performance

Neurocomputing
A wrapper approach for feature selection based on Bat Algorithm and Optimum-Path Forest

Expert Systems with Applications: An International Journal
Comparison between wavelet spectral features and conventional spectral features in detecting yellow rust for winter wheat

Computers and Electronics in Agriculture
Learning algorithms may perform worse with increasing training set size: Algorithm-data incompatibility

Computational Statistics & Data Analysis

Quantified Score

Hi-index	754.84

Visualization

Abstract

The overall mean recognition probability (mean accuracy) of a pattern classifier is calculated and numerically plotted as a function of the pattern measurement complexity n and design data set sizem. Utilized is the well-known probabilistic model of a two-class, discrete-measurement pattern environment (no Gaussian or statistical independence assumptions are made). The minimum-error recognition rule (Bayes) is used, with the unknown pattern environment probabilities estimated from the data relative frequencies. In calculating the mean accuracy over all such environments, only three parameters remain in the final equation:n, m, and the prior probabilityp_{c}of either of the pattern classes. With a fixed design pattern sample, recognition accuracy can first increase as the number of measurements made on a pattern increases, but decay with measurement complexity higher than some optimum value. Graphs of the mean accuracy exhibit both an optimal and a maximum acceptable value ofnfor fixedmandp_{c}. A four-place tabulation of the optimumnand maximum mean accuracy values is given for equally likely classes andmranging from2to1000. The penalty exacted for the generality of the analysis is the use of the mean accuracy itself as a recognizer optimality criterion. Namely, one necessarily always has some particular recognition problem at hand whose Bayes accuracy will be higher or lower than the mean over all recognition problems having fixedn, m, andp_{c}.