Perceptron: Perceptron implementation for pattern recognition
ReTweet Author |
 |
|
Innovation award Nominee: 1x Winner: 1x | ||||||||||
This class implements a model of the Percetron (Artificial Neural Networks - ANN) that can be trained to recognize patterns in its inputs.
This is similar to the algorithm used on palmtops to recognize words written on its pen pad.
The example that comes with this class demonstrates how it can be used to find people that match the profile an inquiring user that fills a form with questions about personal preferences.
First the perceptron is trained with answers of several people. Then the answers of an inquiring user are converted into values that are fed to the perceptron in order to find profile matches.
This is similar to the algorithm used on palmtops to recognize words written on its pen pad.
The example that comes with this class demonstrates how it can be used to find people that match the profile an inquiring user that fills a form with questions about personal preferences.
First the perceptron is trained with answers of several people. Then the answers of an inquiring user are converted into values that are fed to the perceptron in order to find profile matches.
| Classes of André Paterlini Oliveira Vieira | > | Perceptron | > |  Download .zip .tar.gz |
> |  Support forum (1) |
> |  Blog |
> |   Latest changes |
|
|||||||||||||||
| Groups | Applications |  Files |
| Groups |
 Algorithms |
Numerical and statistical algorithms | View top rated classes |
| Artificial intelligence |
Automation of tasks using human-like intelligence | View top rated classes |
| Innovation Award |
 June 2004 Winner Prize: One subscription to the PHP Magazine |
Neural Networking is a computing science field that is not so well known among most software developers, but it can help solving problems that are difficult to solve with conventional computing methods. Neural Networking consists of emulating the way a brain works by the means of networks of interconnected neurons. These neurons have inputs and outputs. They receive input signals that are combined according to some internal factors to define their output signals. The neural networks need to be trained to perform some useful tasks. Usually these tasks consist of perceiving patterns in the input signals of their neurons. This is why they are also known as perceptrons. Training consist of submitting a reasonably large sample of subjects that need to be recognized. Iteratively the values of internal factors of each neuron are adjusted until they output the desired response with an acceptable error rate. The most common applications of neural networks are usually recognizing patterns that otherwise would be hard to distinguish by common computing methods, like for instance, recognizing handwritten characters in a scanned image. This class provides an implementation of a perceptron. It comes with an example application of classifying users by type of profile based on the answers that they provide to questions presented in a Web form. Since the class provides a generic implementation of a perceptron, there could be many other examples of application for this class, possibly only limited by the imagination of the developers that understand its potential. Manuel Lemos |
| Applications that use this package |
No pages of applications that use this class were specified.
If you know an application of this package, send a message to the author to add a link here.
| Files |
| File | Role | Description |
|---|---|---|
 como_funciona |
Doc. | F.A.Q - em português |
| find_match.php |
Example | Example - Finds a class pattern from the submited user profile (see index.html - HTML form) |
| how_does_it_works |
Doc. | F.A.Q. |
 index.html |
Data | Example - RNA Perceptron implementation form |
| perceptron_class.php |
Class | Class - RNA Perceptron OOP class |
 Install with Composer - Download all files: perceptron.tar.gz perceptron.zipNOTICE: if you are using a download manager program like 'GetRight', please Login before trying to download this archive.
|
For more information send a message to
info at phpclasses dot org.Copyright (c) Icontem 1999-2016