Looks like I missed all the action today, was out all day! Jay you're not losing it, none of you are. I knew there would be more questions than answers when I posted the influence response, and having used version 2.1 for 10 minutes I knew it was all gloss, and contained nothing to actually make the nets more efficient. I have tried in the past to manually construct some indicators, as Catlady mentioned, in Excel, but it was more tedious than it was worth, and frankly, the results didn't justify the add'l time and effort. I agree that if Neuro wants to improve it's product (not that there's anything wrong with it now) that they need to add actual preprocessed indicators, increase the training speed, and any of the other great suggestions that have been made on this thread during the last few months... things that will help us profit while reducing our risk.
I for one will be making my suggestions to Neuro as requested, and I hope you all do to. We have an opportunity in helping to develope what could be one of the best products of its kind on the market.
Now some questions that I have regarding the response... support stated "influence period is simply the number of days that NS will look back
when making a prediction". Does this mean that the influence period does,'t influence the training cycle, just the predictions? It seems so.
Then they stated "so NS can pickup most signals if the influence period is long enough." If that's the case why not set all the influences for 30 days? Also it seems to indicate that not only does Neuro look back the specified influence days but it picks up meaningful relationships during the entire duration of the set period.
Frankly I think I'm as confused as I was before the response.
Happy Jay? Now I've loooossstt it too! |
