This is the NeuroDimension Newsletter, which you are receiving because you requested to stay informed about new developments at NeuroDimension. If you would like to stop receiving these newsletters, please see the bottom of this newsletter for instructions.
In this issue...
|
Neural Network Course
Completed! Customer Rating – 
The October Neural Network Course was a great success! We at
NeuroDimension would like to thank all of the attendees for all their great questions and feedback on the course. We always enjoy getting to know our customers better, especially when it gives us
a chance to discuss their applications. Once again, we received excellent reviews from our attendees and have never
had anyone tell us they would not recommend our course to others.
|
|
The next Neural Network course is being planned for Spring 2005. If you would like more information on our courses, please visit:
http://www.neurosolutions.com/products/course/
|
TradingSolutions v3.0 Now Available!
NeuroDimension is pleased to announce the release of TradingSolutions v3.0. Many exciting new features have been added to the software, all combining to simplify the process of creating profitable predictions. Many improvements have also been made to the interface, simplifying the process of creating predictions. Furthermore, the capabilities and potential uses of genetic optimization have been significantly improved.
|
|
Major Features added in TradingSolutions v3.0
- Optimization of Predictions Based on Trading Performance
- Optimization of Prediction Input Parameters
- Optimization of Entry/Exit System Parameters
- Optimization of Voting in Committees of Signals
- Optimization of Group Fields
- Video Help System
- Improved User Tutorials
TradingSolutions v3.0 also includes the ability to write state-of-trade exits in entry/exit systems using the user friendly point-and-click function editor. This allows you to perform exits on trade length, profit, loss, and many other characteristics.
Evaluate TradingSolutions v3.0 for yourself!
A free evaluation copy of TradingSolutions v3.0 is available for download from www.tradingsolutions.com. Included with the 30-day evaluation is the Solution Service. The Solution Service provides users with daily trading signals from several of NeuroDimension's top-performing neural network models, as well as stock data and commentary for those models. A few of the Solution Service stocks include Starbucks (NASDAQ:SBUX), Microchip Technologies (NASDAQ:MCHP), Baxter International (NYSE:BAX), and Ford Motor Company (NYSE:F).
For complete information on new features, evaluating and ordering please visit: http://www.tradingsolutions.com/resources/whatsnew.html
|
|
Building a LVQ Network within NeuroSolutions
A competitive network learns to classify input vectors based solely on the distance between those vectors. If two input vectors are very similar, the competitive network will likely assign them to the same class. Since competitive networks are unsupervised, there is no way to assign input vectors to target classes.
|
|
Learning vector quantization (LVQ) is a method for training competitive networks in a supervised manner. In other words, they learn to classify input vectors into target classes. The concept behind the LVQ algorithm is quite simple:
- If the competitive rule produces the correct class label, then no modification to competitive learning is necessary.
- If the competitive rule produces an incorrect class label, then the PE weights should be "repelled" from the present cluster.
For more details on this algorithm, please refer to section 7.8 of the Interactive Book or Kohonen, T. Self-Organization and Associative Memory, 2nd Edition, Berlin: Springer-Verlag, 1987.
To implement LVQ in NeuroSolutions you start with a Competitive network and then add a File component to the "Forced" access point of the Synapse. The file you specify provides the target classes (i.e., the desired output). In addition to the examples shown in the NeuroSolutions Demos (under "Unsupervised Learning") and the Interactive Book (Example 7.12), we have put together a simple classification example using LVQ and made it available at:
http://www.nd.com/public/support/LVQExample.zip
This network solves the same "Crab Classification" problem that is presented within the Example sections of the NeuroSolutions Getting Started Manual (under the Help menu). You will see that it does a reasonable job of classification, but not as good as an MLP. There may be some applications where a LVQ network will produce a better (i.e., more general) solution than a MLP since there are fewer network weights. There are also more sophisticated variants of the LVQ (e.g., LVQ2, LVQ3, etc.) that can be implemented with user-defined components (i.e., DLLs written in C).
|
Customer Spotlight: Texture Classification – Electronic nose and neural network use for the classification of honey
Simona BENEDETTI, Saverio MANNINO, Anna Gloria SABATINI, Gian Luigi MARCAZZAN
Abstract
|
|
Seventy samples of honey of different geographical and botanical origin were analyzed with an electronic nose. The instrument, equipped with 10 Metal Oxide Semiconductor Field Effect Transistors MOSFET) and 12 Metal Oxide Semiconductor (MOS) sensors, was used to generate a pattern of the volatile compounds present in the honey samples. Principal Component Analysis (PCA) and Artificial Neural Network (ANN) evaluated the sensor responses. Good results were obtained in the classification of honey samples by using a neural network model based on a multilayer perceptron that learned using a backpropagation algorithm. The methodology is simple, rapid and results suggest that the electronic nose could be a useful tool for the characterization and control of honey.
Complete application summary available at:
http://www.neurosolutions.com/resources/apps/honey.pdf
|
|
TradingSolutions Live is an introductory online course in which a standard web-browser is used to view the instructor's computer screen.
The course provides a broad overview of TradingSolutions v3.0 covering everything from creating a “traditional” rule-based trading system to creating a powerful neural network trading system. Neural network systems have a distinct advantage over rule-based systems in that they learn patterns directly from historical data, allowing you to create systems with more accurate signals for entering and exiting positions.
More information and to sign up for a course, visit: http://www.tradingsolutions.com/resources/tslive.html
|
|
We appreciate your feedback! Please send us your comments or suggestions concerning this newsletter, our web site, or part of the NeuroDimension product line. Write to us at: newsletters@nd.com.
Have questions about NeuroDimension products? Send your questions to: info@nd.com
This issue and previous issues of this newsletter are available on the NeuroSolutions web site at: http://www.neurosolutions.com/newsletters.html
If you would prefer not to receive these newsletters or subsequent product updates from NeuroDimension, please reply to this letter with the subject heading changed to the word REMOVE.
You can also contact us with questions or removal requests via mail at:
NeuroDimension, Inc., 3701 NW 40th Terrace, Suite 1, Gainesville, FL 32606
Thank you again for your support of NeuroDimension products!
|
|
