What type of software is available for these non-linear systems? How do various packages compare and at what cost? What are the strengths & weaknesses of these black-box approaches? How much of a learning curve is required? Let's start a thread here!
Back to the top again!
HF....This is a post I sent awhile ago, unfortunately with no response.........
"Perhaps this section of the T2W BB is not the best place to post this e-mail but........
Hill Farmer posed a question of Neural Networks on the 15th and there was little response. So, any T2W member had any experience as to the effectiveness of DeepInsight especially since a new 2001 version has come out? Anyone "bent" current data to fit the limitations of the free trial and what was their conclusion.
Info available at http://www.deepinsight.com if anyone wishes to conduct any experiments.
I have absolutely no associations with the company but posting this purely to for the benefit of T2W members.
rgds to all...............
Whilst DEEPINsight is not strictly a Neural Network but rather a self-training system using standard indicators which seem to measure the correlation error to adjust themselves.
You may wish to experiment; I will when I get time. I would appreciate your findings.