AmiBroker eSignal Extreme Charts Generic DLL Investor R/T Market Delta MATLAB MetaStock Microsoft Excel MultiCharts NeoTicker NeuroShell NinjaTrader Tradecision Trade Navigator Trading Solutions TradeStation

Create leading indicators

Create complex concurrent indicators

Apply to stocks, futures, ... almost anything

Neural networks are mathematical functions that you "train" to covert input data to whatever you specify as output data. For example, you may want to convert various technical indicators into:

• leading indicators.
• an assessment of whether or not a market reversal is imminent.
• a forecast of the MACD indicator a few bars into the future.

You may not know the math to perform such calculations, but a Neural Net may be able to figure it out! Afterwards, you can develop trading decisions based on the neural net's forecasts.

The power of Braincel is well established. In each of the following articles in Future's Magazine, Braincel was used to build a leading indicator...

• Futures Aug '94Training Neural Nets for Intermarket Analysis
• Futures May '96How to Predict Tomorrow's Indicators Today
• Futures Dec '96Nets and Intermarket Divergence

BRAINCEL is an Excel add-in that enhances your forecasts with the power of neural networks. Amazingly, no knowledge of neural net math or statistics is required to use it. You supply both input data and desired output (target) data in rows and columns on a spreadsheet. Braincel will study this data and after training, you supply it with new input data and Braincel will give you answers that will astound you! It's almost as simple to use as Excel's standard regression tool, but more powerful!

To create a neural net, you simply set up your data on an Excel spreadsheet, select a few options, and the neural net will automatically train and test itself. The neural net models you produce will work in ...

• Microsoft Excel
• TradeStation 2000i, 6, 7, 8
• your own custom programming in C and VB.

Novices enjoy the guidance from standard dialog boxes and programmers love the power in using Excel's VBA to automate the more complex tasks that Braincel can perform.

Braincel's advanced features include . . .

• Jurik's own algorithms. It does not use the old Back-Propagation formula
• Built-in Excel charts show training progress
• Tables for total control over learning parameters
• Automatic search for the best net architecture
• Automatic search for the best selection of inputs

Here's how to create a neural net for use in TradeStation 2000i, 6, 7 or 8...

1. Plot your favorite technical indicators on your TradeStation screen.
2. Export the chart data to a .CSV file and open the file with Microsoft Excel.
3. Create and train a neural net on the spreadsheet data and save it.
4. Write code in Easy Language to apply the neural net to market data. You can access your neural net within INDICATORS, SYSTEMS, and USER FUNCTIONS. (Example Easy Language code is provided.)
5. Run your new studies on real-time or end-of-day data!

For a working example of how you can access from TradeStation Easy Language, a neural net built by Braincel, look for "Braincel - Neural Net Example" in our "Free Stuff" section.

To run Braincel and create neural nets, you need one of the following:

You can run your newly created neural nets on market data in Microsoft EXCEL and TradeStation

LATEST VERSION: Braincel 3.7released JUNE 2005

To determine if you have the latest version, proceed as follows:

1. Open the folder on your hard drive containing the installed Braincel files
2. Right click on the file "BRAINCEL.XLL" and select PROPERTIES
3. For version 3.7, the "Created" date should read 27-MAY-2005.
4. To order the latest version, go to our SHOPPING CART and select the "Upgrade/Replace" department.

Download our TECHNICAL REPORTS. Lots of charts and comparisons. Read our CUSTOMER's LETTERS. See what others have to say. Read our FREQUENTLY ASKED QUESTIONS. For answers to technical issues. Prices, order form, specific questions to ask.

SPECIAL REPORT: BACK-PERCOLATION

Braincel uses a newer algorithm called Back-Percolation instead of the rather old formula called Back-Propagation. Back-Percolation was designed by Mark Jurik and this report explains his philosophy of the successful algorithm. For a real visual treat, the report displays parameter learning trajectories of amazing symmetry and beauty. --- Go to TECHNICAL REPORTS page.

[ Copyright ] [ WebMaster ]