Letter recognition using Holland-style adaptive classifiers

By Peter W. Frey, David J. Slate. 2004

400 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

BOHB: Robust and Efficient Hyperparameter Optimization at Scale (2018, background): “We optimized six hyperparameters that control the training procedure (initial learning rate, batch size, dropout, exponential decay factor for learning rate) and the architecture (number of layers, units per layer) of a feed forward neural network for six different datasets gathered from OpenML (Van”
Recent Advances in Open Set Recognition: A Survey (2018, background): “datasets and their combinations: LETTER [135]: has a total of 20000 samples from 26 classes, where each class has around 769 samples with 16 features.”
Multiple instance learning: A survey of problem characteristics and applications (2016, methodology): “Letters [148]: This semi-synthetic MIL data set uses instances from the Letter Recognition data set2.”
Multivariate Density Estimation by Bayesian Sequential Partitioning (2013, result): “According to the previous reports on this dataset, HollandStyle Adaptive Classifiers achieve a 82% correct classification rate (Frey and Slate 1991), while Mixtures of Conditional Maximum Entropy Models achieve 76.”
Multivariate decision trees using linear discriminants and tabu search (2003, neutral): “Letter. This letter recognition data was studied in [ 14 ] and the StatLog project.”
Learning classifier systems: a complete introduction, review, and roadmap (2009, positive): “Although not a new idea [3, 25, 166], the accuracy-based fitness of XCS represents the starting point for a new family of LCSs, termed “accuracy-based” which are distinctly separable from the family of “strengthbased” LCSs epitomized by ZCS (see Table 1).”
Deep Neural Decision Forests (2015, neutral): “G50c [33] Letter [10] USPS [14] MNIST [20] Char74k[8] Tree input features 10 (random) 8 (random) 10x10 patches 15x15 patches 10 (random) Depth 5 10 10 10 12 Number of trees 50 70 100 80 200 Batch size 25 500 250 1000 1000 Table 1 .”
Accuracy-Based Learning Classifier Systems: Models, Analysis and Applications to Classification Tasks (2003, methodology): “The accuracy computation is equivalent to Frey&Slate’s classifier system (Frey and Slate, 1991).”
A Constructive Hybrid Structure Optimization Methodology for Radial Basis Probabilistic Neural Networks (2008, methodology): “The best test accuracy obtained by Frey and Slate [ 28 ] using other methods is only a little over 80%, which is much lower than the test accuracies obtained by the RBFNN and the RBPNN.”
K-means properties on six clustering benchmark datasets (2018, background): “The Letter dataset [41] contains 26 classes representing letters from A to Z.”
Extreme Value Theory for Open Set Classification - GPD and GEV Classifiers (2018, methodology): “This protocol is based on the LETTER data set [26] that contains a total of 20000 observations of 26 different classes corresponding to handwritten letters.”
The Extreme Value Machine (2015, methodology): “was obtained by a 3-fold cross-validation over the LETTER training set [35], choosing k from {1, 4, 10}.”
Probability Models for Open Set Recognition (2014, methodology): “For evaluation, we breathe new life into the classic data sets for multi-class classification, LETTER [12] and MNIST [21], by changing their testing protocols.”
End-to-End Learning of Decision Trees and Forests (2019, methodology): “Table 2 Properties of the datasets used in our decision forest experiments, respectively published by Frey and Slate (1991), Hull (1994 The lower part describes our processing of the data.”