ForeTrade programs and files currently available for download:
Note: Our NeuroShell Trader indicators are no longer available for download.
Please read our Disclaimer before downloading any of our files below.
All our programs are virus free.
The toolset cannot be downloaded directly.
Please contact us with your TS client number. A dedicated trial version (DLL, Indicators and sample workspaces) will be compiled for your account.
Reminder: The sample files below are only there to demonstrate GPF output. Not only it cannot show the full potential of the GPF, they are also not meant display optimal signals. The best signals can only be the ones tailor made with your own GPF settings.
Click box below to download:
GPF is now free. If you do not yet have installed the unconstrained version, please contact me and i'll send you a permanent set of codes.
4. The provided samples have been kept very simple, therefore cannot show all features from the G.P.F. You should have no trouble finding better patterns yourself on your favorite stocks. The Cisco file is commented. They are posted here for tutorial purposes, and data is not up-to-date.
6. Upgraders should first save their new settings to update the GPF registry settings. There may be new features that need to be reset in the Windows Registry. Should the GPF behave erratically, please delete your Registry settings ("Other Settings" tab), and save them again.
7. Besides additional features, the G.P.F. Pro also includes the Omega and NeuroShell Trader interfaces. These interfaces are not available on line, as there is no need for a trial version, and they do not require user registration. Please contact us for any information you may need. They will be directly emailed to you on request.
This VB Neural Net program is again a Excel Add-in. The (beta) final version is available below, as well as the reduced shareware version. Please contact us at firstname.lastname@example.org if you are interested in acquiring the VB source code. This program is an excellent neural net tutorial for VB programmers, but is not used in our developments, so will now remain as is. For larger neural projects, we recommend you to rather consider NXL (see below).
The VB BackProp requires a validation code to be enabled. You will find your own file ID on the 'ForeTrade' tab. Please email it to email@example.com Upon payment (see our sales page) ou'll receive a validation number, which will permanently unlock the program. The demo version is limited to 100 patterns and 5 inputs.
We wish to remind you that it remains a developer's tool, and there is no better way to make the most of it than buying the source code, and making it your own neural net for MS Excel or MS Access.
The VB Backprop zipped file does include the shareware version. The update only contains the latest update XLA file for a faster download.
See our Neural Net page for more details.
In this second release, NXL2 is a fast neural net library close to QuickProp (Prof Scott Fahlman), easily called from Microsoft Excel, or other VB/VBA clients like MS Access. The neural library is written in C for optimal performance. The algorithm has been streamlined, and uses optional connection pruning to keep the neural architecture as simple as possible.
The DLL will run in demo mode (with some neural network size limitations) until unlocked with a personal code. The small DLL unlocking utility is provided in the same ZIP file.
Cascade-correlation has not been implemented yet. Finding the optimal number of number of hidden neurons will be possible in a future GNXL (Genetic NXL).
QP2.2 and QP3.2 examples are similar function mapping experimentations. Classif_NXL, as its name implies is a classification exercise. AOL NXL is a simple example of prediction using real AOL data. Inputs are not optimized though.
The utility is considerably faster, and more versatile than the original wavelet spreadsheet from Prof. Murtagh. The trade-off to an extent, is size. The compressed version is around 230KB (~800KB uncompressed).
Some knowledge of the wavelet concept is required to run this utility. A sound understanding of Prof. Murtagh's work is certainly preferable (c.f. www.multiresolution.com). This utility is a rewrite of Prof. Murtagh a-trous wavelet version 2, now replaced and therefore apparently no longer available on the MultiResolution site.
There has been much debate on the use of wavelets as neural net inputs. Your feedback will be most appreciated at firstname.lastname@example.org
The equivalent NSTrader Wavelet Add-on has been rewritten, thanks to Victor Chu's contribution, and is now also available below. Also exists for TradeStation on www.tradestationworld.com.
NOTE: This work only represents a one-off contribution to a-trous wavelets, which may be dating a bit now. We have not pursued this type of wavelet any further.
WinSOM has been developed in Borland C++ (BCB5). Based on the Kohonen SOM algorithm, its main purpose is to reduce dimensionality in clustering data onto a 2D grid or cluster map. For instance, if one commonly uses 5 or 10 or more (uncorrelated) indicators (such as Forward PE, Sharpe, correlations to various markets, etc,...) it will "naturally" group patterns having similar features. Trivial visual examples are included to show the benefits of using SOM. It is relatively easy to use, but some prior knowledge of the Kohonen SOM technique is recommended.
Please note that WinSOM has not been tested with all locales of Windows. It will run fine on US-English 32-bit versions of MS Windows. If you have any trouble reading pattern files (in text format), please take cue from the provided data sample using Windows NotePad.
This product is now free but... 100% unsupported.
An Excel pattern file generator is included in the WinSOM download.
Please note that the C++ source code is available for sale (Borland C++ Builder 5). Contact us for details.
No bugs reported. Please check for patches on the WinSOM page. Current version in V1.2
Details on the IB_Direct page
There are tons or freeware or shareware available on the Internet. Here will be listed only those including C/C++/VB source code.
Please note that the Fibo and TimeSynch executables require the MS.NET framework: http://msdn.microsoft.com/netframework/downloads/howtoget.aspx
The first small utility displays a very simple example of chaotic evolution based on the well-known logistic function. It is free to download and copy, but is not supported. You may want to search newsgroups such as sci.nonlinear for information on the chaos theory. The book 'Chaos' from James Gleick is also a good non technical introduction to the topic.
The second is particularly simple and utterly useless.
Thank you for your support.
Page last modified:
December 09, 2007