network profile
| bio | website | interruptsoftware.ca |
|---|---|---|
| location | ||
| age | ||
| visits | member for | 10 months |
| seen | May 15 at 0:57 | |
| stats | profile views | 8 |
Technologist, Entrepreneur & Artist
|
May 14 |
comment |
Grokking Stochastic Oscillator for Stocks Point taken. I've just programmed the %D, and seem to be getting reasonable numbers. Seeing those rapid changes makes sense - thanks for that. As far as the days closing price, we're essentially just dealing with a time series. So whether it's weekly, daily, sub-second, etc, I expect an algo such as this to nominally tell me the same information. Over the next few days, I want to put it into a graph, so I can see if it makes sense. I'll keep you posted. Cheers. |
|
May 14 |
awarded | Commentator |
|
May 14 |
comment |
Grokking Stochastic Oscillator for Stocks I'm using second and sub-second ticks. Ie, over the last 14 ticks, I would take the High and Low, and the last tick's closing trade. So what you're saying is that my 14 ticks, would have an exponential smoothing over the first 3 of those ticks. Seems like a short time for smoothing. But I can try it out. |
|
May 14 |
comment |
Grokking Stochastic Oscillator for Stocks I dont think it matters, whether or not, ticks are evenly distributed over time. %K is the last tick's percentage move within the high / low range (over, say, 14 ticks - see here). %D is the 3-period exponential moving average of %K. I want to know what that period means. Is it the last 14 ticks, 3 times? Is it only the last 3 periods within the last 3 ticks (say, 6 secs)? Remember, %D is a smoothing factor, so I don't see that time period being long enough. And I dont think it matters how those ticks are distributed over time. |
|
May 14 |
comment |
Grokking Stochastic Oscillator for Stocks But let's say each tick happens every 2 seconds. If the %D period is the time between two ticks, then the 3 would represent all ticks that occurred within the last 6 (3 * 2 seconds) seconds. That i) obviously isn't enough time for smoothing, and ii) doesn't correspond to your 3/14 of a day. |
|
May 13 |
asked | Grokking Stochastic Oscillator for Stocks |
|
Nov 4 |
comment |
Multilayer Perceptron (Neural Network) for Time Series Prediction Ok, so for this problem, I've gotten further along in building a NN and it's training (using backprop). But I'm still getting some funny results in my weight updates, and consequently, the prediction targets. I've outlined the problem more fully on the Clojure mailing list (click here). Any insights are appreciated. |
|
Nov 4 |
comment |
Howto Calculate An Error's Partial Derivative in ANN Ok, for the backprop algorithm, I've gotten a bit further. But I'm still getting some funny results in my weight updates, and consequently, the prediction targets. I've outlined the problem more fully on the Clojure mailing list click here. Any insights are appreciated. |
|
Oct 8 |
comment |
Howto Calculate An Error's Partial Derivative in ANN Yeah, I'm starting to get the jist of from the paper you provided - it's a doozey. I'm on p.16, "Steps of the algorithm". I'm sure I'll have 1 or 2 more Qs. It really comes down to the implementation details. As I'm trying to use this to build a neural network (see here) |
|
Oct 5 |
comment |
Howto Calculate An Error's Partial Derivative in ANN Ahh, I'm getting what you mean by that statement. Starting to sink in slowly :) |
|
Oct 5 |
comment |
Howto Calculate An Error's Partial Derivative in ANN I conceptually understand what you mean by "the gradient of any individual weight depends on weights and activations appearing later in the network". But I don't understand how that bears out when calculating weight changes. I've been using this paper as a guide to backprop. Anyways, I'll definitely go through the paper you provided. Thanks. |
|
Oct 5 |
comment |
Howto Calculate An Error's Partial Derivative in ANN Thanks for this. Andrew Ng is actually the guy whose videos I first started following, when I was getting into AI. But this course seems to emphasize using NNs for classification problems (discreet values). I want to compute a continuous value in a time series. But I will dig further into this series. Thanks again. |
|
Sep 30 |
asked | Howto Calculate An Error's Partial Derivative in ANN |
|
Jul 16 |
comment |
Multilayer Perceptron (Neural Network) for Time Series Prediction Thanks very much for your clarity. A) In the output layer, how would you derive the values in Z(t), a vector, if the hidden layer is firing to you, a single value, Zn(t) ? B) Using Back Propagation, to adjust the weights and biases. I'm assuming we would use gradient descent to adjust those weights? How is the bias adjusted? Or is that just the threshold that the user supplies? C) So you would recommend Time Delayed Neural Networks over using a sliding window input or running averages as inputs to my Multilayer Perceptron. Thanks again. |
|
Jul 15 |
awarded | Student |
|
Jul 15 |
asked | Multilayer Perceptron (Neural Network) for Time Series Prediction |
