Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

CARVE--a constructive algorithm for real-valued examples.

S Young¹, T Downs

¹Department of Experimental Psychology, University of Oxford, OX1 3UD, UK.

IEEE Transactions on Neural Networks

|February 8, 2008

Summary

This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Neonatal intensive care unit discharge preparation, family readiness and infant outcomes: connecting the dots.

Journal of perinatology : official journal of the California Perinatal Association·2013

Same author

A systematic review of integrative oncology programs.

Current oncology (Toronto, Ont.)·2013

Same author

Thyrotrophin receptor antibody characteristics in a woman with long-standing Hashimoto's who developed Graves' disease and pretibial myxoedema.

Clinical endocrinology·2012

Same author

A study of venous thrombosis incidence in patients with acute hyperthyroidism.

Internal medicine journal·2012

Same author

Antibiotic timing at caesarean section.

International journal of obstetric anesthesia·2012

Same author

EGFR tyrosine kinase mutation testing in the treatment of non-small-cell lung cancer.

Current oncology (Toronto, Ont.)·2012

Same journal

Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations.

IEEE transactions on neural networks·2013

Same journal

Guest editorial: special section on white box nonlinear prediction models.

IEEE transactions on neural networks·2011

Same journal

Data-based fault-tolerant control of high-speed trains with traction/braking notch nonlinearities and actuator failures.

IEEE transactions on neural networks·2011

Same journal

Guest editorial: special section on data-based control, modeling, and optimization.

IEEE transactions on neural networks·2011

Same journal

Neural network-based multiple robot simultaneous localization and mapping.

IEEE transactions on neural networks·2011

Same journal

Data-driven model-free adaptive control for a class of MIMO nonlinear discrete-time systems.

IEEE transactions on neural networks·2011

See all related articles

A new constructive neural network algorithm, CARVE, efficiently creates feedforward networks for real-valued classification tasks. It offers near-minimal solutions and can initialize other training schemes like backpropagation.

Area of Science:

Machine Learning
Artificial Intelligence
Computational Neuroscience

Background:

Neural networks are powerful tools for classification tasks.
Existing constructive algorithms often struggle with real-valued input data.
Efficient training and network minimization remain key challenges in neural network research.

Purpose of the Study:

To present a novel constructive neural network algorithm, CARVE.
To extend sequential learning to handle real-valued input vectors.
To develop an efficient method for constructing near-minimal feedforward networks.

Main Methods:

The CARVE algorithm constructs feedforward networks with a single hidden layer of threshold units.
It extends the sequential learning algorithm to real-valued inputs.

Related Experiment Videos

Convex hull methods are employed for determining network weights.

Main Results:

CARVE successfully implements consistent classification tasks on real-valued data.
The algorithm was applied to benchmark datasets like sonar, Monks, and iris data.
It demonstrated efficiency in producing near-minimal network solutions.

Conclusions:

CARVE provides an efficient constructive approach for neural network training.
The algorithm effectively handles real-valued classification tasks.
CARVE serves as a valuable tool for initializing network topologies and weights for other training methods, such as backpropagation.