Bayesian selection of important features for feedforward neural networks

Priddy, Kevin L.; Rogers, Steven K.; Ruck, Dennis W.; Tarr, Gregory L.; Kabrisky, Matthew

doi:10.1016/0925-2312(93)90030-7

Cited by 52 publications

(19 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [12], [24] and [25], the absolute values of the weights in the first hidden layer are used to that end. The derivative of the output function with respect to the input features has also been widely used to compute their saliency, as in [14], [17], or [19].…”

Section: Classification and Feature Selection Using An Artificial Neumentioning

confidence: 99%

Classification, Dimensionality Reduction, and Maximally Discriminatory Visualization of a Multicentre 1H-MRS Database of Brain Tumors

Lisboa

Romero

Vellido

et al. 2008

2008 Seventh International Conference on Machine Learning and Applications

View full text Add to dashboard Cite

show abstract

Section: Classification and Feature Selection Using An Artificial Neumentioning

confidence: 99%

Classification, Dimensionality Reduction, and Maximally Discriminatory Visualization of a Multicentre 1H-MRS Database of Brain Tumors

Lisboa

Romero

Vellido

et al. 2008

2008 Seventh International Conference on Machine Learning and Applications

View full text Add to dashboard Cite

show abstract

“…Eq. (1) exemplifies such a measure [25,26] (1) where y is the predictor output, Q is the number of outputs, P is the number of training samples, and x ip is the i th component of the p th input vector x p .…”

Section: Introductionmentioning

confidence: 99%

“…Torkkola and Tuv suggest using artificial-contrast variables created by randomly permuting values of original N variables across the data points [29]. Some of feature selection procedures are based on making comparisons between the saliency of the candidate and the noise feature [25,26]. One of the main drawbacks of the feature saliency measures is that the measures do not have direct relation to the prediction error.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A feature selection technique for generation of classification committees and its application to categorization of laryngeal images

Bačauskienė

Verikas

Gelžinis

et al. 2009

Pattern Recognition

View full text Add to dashboard Cite

This paper is concerned with a two phase procedure to select salient features (variables) for classification committees. Both filter and wrapper approaches to feature selection are combined in this work. In the first phase, definitely redundant features are eliminated based on the paired t-test. The test compares the saliency of the candidate and the noise features. In the second phase, the genetic search is employed. The search integrates the steps of training, aggregation of committee members, selection of hyperparameters, and selection of salient features into the same learning process. A small number of genetic iterations needed to find a solution is the characteristic feature of the genetic search procedure developed. The experimental tests performed on five real-world problems have shown that significant improvements in classification accuracy can be obtained in a small number of iterations if compared to the case of using all the features available.

show abstract

“…Also the problems of feature selection is dealt in. [2] [11][12][13][14][15][16][17]. Principal component analysis (PCA) is one of the famous method used [11].…”

Section: Introductionmentioning

confidence: 99%

Feature Selection by Mining Optimized Association Rules based on Apriori Algorithm

Rajeswari¹

2015

IJCA

View full text Add to dashboard Cite

This paper presents a novel feature selection based on association rule mining using reduced dataset. The key idea of the proposed work is to find closely related features using association rule mining method. Apriori algorithm is used to find closely related attributes using support and confidence measures. From closely related attributes a number of association rules are mined. Among these rules, only few related with the desirable class label are needed for classification. We have implemented a novel technique to reduce the number of rules generated using reduced data set thereby improving the performance of Association Rule Mining (ARM) algorithm. Experimental results of proposed algorithm on datasets from standard university of California, Irvine (UCI) demonstrate that our algorithm is able to classify accurately with minimal attribute set when compared with other feature selection algorithms.

show abstract

Bayesian selection of important features for feedforward neural networks

Cited by 52 publications

References 5 publications

Classification, Dimensionality Reduction, and Maximally Discriminatory Visualization of a Multicentre 1H-MRS Database of Brain Tumors

Classification, Dimensionality Reduction, and Maximally Discriminatory Visualization of a Multicentre 1H-MRS Database of Brain Tumors

A feature selection technique for generation of classification committees and its application to categorization of laryngeal images

Feature Selection by Mining Optimized Association Rules based on Apriori Algorithm

Contact Info

Product

Resources

About