Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Classification trees with neural network feature extraction.

H Guo1, S B Gelfand

  • 1Sch. of Electr. Eng., Purdue Univ., West Lafayette, IN.

IEEE Transactions on Neural Networks
|January 1, 1992
PubMed
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

Hormone replacement therapies, oral contraceptives, reproductive factors and colorectal adenoma risk: a systematic review and dose-response meta-analysis of observational studies.

Colorectal disease : the official journal of the Association of Coloproctology of Great Britain and Ireland·2019
Same author

The immune profile of small HER2-positive breast cancers: a secondary analysis from the APT trial.

Annals of oncology : official journal of the European Society for Medical Oncology·2019
Same author

First Report of Phytophthora palmivora Causing Root Rot of Cassava in China.

Plant disease·2019
Same author

Peripheral neutrophil CD64 index combined with complement, CRP, WBC count and B cells improves the ability of diagnosing bacterial infection in SLE.

Lupus·2019
Same author

Expression patterns and colocalization of two sensory neurone membrane proteins in Ectropis obliqua Prout, a geometrid moth pest that uses Type-II sex pheromones.

Insect molecular biology·2018
Same author

[Clinicpathologic studies of cases with tracheobronchopathia osteochondroplastica].

Zhonghua bing li xue za zhi = Chinese journal of pathology·2018
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

This study proposes using small multilayer neural networks within classification trees to extract nonlinear features. This hybrid approach significantly reduces error rates and tree size compared to traditional methods.

Area of Science:

  • Machine Learning
  • Artificial Intelligence
  • Computer Science

Background:

  • Standard classification tree methods can struggle with complex nonlinear feature extraction.
  • Training large, unstructured neural networks presents significant computational challenges.
  • Selecting optimal neural network architecture (size) is a critical and difficult task.

Purpose of the Study:

  • To propose an integrated approach combining small multilayer neural networks with classification trees.
  • To enhance nonlinear feature extraction within decision nodes of binary classification trees.
  • To improve classification accuracy and reduce model complexity.

Main Methods:

  • Utilizing small multilayer neural networks at decision nodes for nonlinear feature extraction.

Related Experiment Videos

  • Employing a gradient-type learning algorithm for training nets and growing the tree in multiclass scenarios.
  • Developing an efficient tree pruning algorithm to optimize tree size and complexity.
  • Comparing the proposed method against the CART (Classification and Regression Trees) method and standard backpropagation.

    • Main Results:

    • The proposed method consistently yields trees with lower error rates and fewer nodes.
    • Significant decreases in both error rate and tree size were observed on waveform and handwritten character recognition tasks.
    • The approach effectively reduces issues associated with training large, unstructured neural networks.
    • Comparable error rates and shorter training times were achieved compared to large multilayer nets trained with backpropagation.

    Conclusions:

    • Integrating small multilayer neural networks into classification trees offers a superior method for nonlinear feature extraction.
    • This hybrid model provides a more efficient and effective alternative to traditional classification trees and large standalone neural networks.
    • The proposed method demonstrates improved performance in terms of accuracy and model parsimony for pattern recognition tasks.