Entropy nets: from decision trees to neural networks

Article Abstract:

Decision trees can be restructured as multiple-layer artificial neural networks (ANNs). ANNs are composed of many interconnected neurons. The neurons are arranged in layers; neurons on the network's input side compose the sensory layer, while the neurons on the output side compose the motor control layer. All layers between the sensory and motor control layers are called hidden layers. Decision trees can be restructured as three-layer networks, called entropy networks. Entropy nets are advantageous because they require fewer neural connections than other networks. The reduction of neural connections is advantageous in chip design. Entropy nets can be mapped directly from decision trees, which can be generated via data-driven procedures.

Pattern recognition (Computers), Neural networks, Decision trees, Pattern Recognition, Computer Learning, Methods, Classification Systems, Decision Tree, Entropy, Neural Network

---

Modern permanent magnets for applications in electro-technology

Article Abstract:

Uses for magnets are expanding as traditional devices become miniaturized and new applications and design concepts evolve. The three major commercial magnet materials are the Alnico Magnet family, the ferrite or ceramic magnetics; and the rare-earth magnet family. Permanent magnets are used in electric motors, generators, electro-mechanical actuators, measuring instruments, electric current control, sound generators, sound receivers, mechanical force and torque applications, power tubes, waveguide devices, particle accelerators, synchrotron radiation sources, free electron lasers, mass spectrometers, cathode ray tubes, transducers, and medical electronics and bioengineering.

Machinery, Electronics, Magnetic devices, IEEE

---

Similar abstracts:

• Abstracts: Second-order stochastic properties in queueing systems. Petri nets: properties, analysis and applications
• Abstracts: The electrical properties and device applications of homoepitaxial and polycrystalline diamond films
• Abstracts: New flow cytometric capabilities at the National Flow Cytometry Resource. Design considerations for achieving MLS Category III requirements
• Abstracts: Radar signal categorization using a neural network. Sensitivity Analysis of the Ampacities of Self-Contained, Oil- Paper-Insulated, Naturally Cooled, and Forced Cooled 300-kV Underground Cables
• Abstracts: VSAT data networks. Rationale, directions, and issues surrounding high speed networks

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.