Using neural networks in trading - page 38 1...31323334353637383940 New comment Aleksei Morozov 2013.07.26 07:34 #371 A question for NeuroSolutions users. Has anyone used Kohonen? Any of the following: LineK, SquareK, DiamondK. Or StandardFull, ConscienceFull. Vizard 2013.07.26 07:46 #372 Demi: so what are the mystery pictures? there's nothing mysterious there... solar 2013.07.26 18:15 #373 Demi: so what's with the mystery pictures? The currency indexes, there is nothing mysterious about them. ))) [Deleted] 2013.07.26 18:21 #374 solar: The indices are currency indices, there's nothing mysterious about them. ))) and what do they show? Vladimir Gomonov 2013.07.26 18:25 #375 FAGOTT: and what do they show? A thick provocation towards the Stat Arbitrage. [Deleted] 2013.07.26 18:28 #376 I have visions on Friday nights, too. with an hors d'oeuvre, with an hors d'oeuvre! solar 2013.07.26 18:32 #377 FAGOTT: I have visions on Friday nights, too. with an hors d'oeuvre, with an hors d'oeuvre! In such cases, disulfiram can help. Or buprenorphine. [Deleted] 2013.07.26 18:34 #378 solar: It's like looking at )))) The blue one is the euro, the green one is the quid. Eurobucks chart . I will not write a clever article. You can see it all. What and why? Дмитрий 2013.10.20 17:11 #379 Hi there. Has anyone tried the following type of construction: The first Kohonen map is trained, let's call it SOM_in. It will pre-process the input data. A second Kohonen map is trained called SOM_out. It is trained to classify the output examples. Learning perseptron in the input it has cluster numbers from SOM_in and the response is a cluster from SOM_out. And another perseptron is trained to "decode" the cluster number of SOM_out into the final prediction data. ??? Kohonen and Patterns Discussion of article "Using Market etiquette or good Mihail Marchukajtes 2013.10.21 00:46 #380 TimeMaster: Hi there. Has anyone tried the following type of construction: The first Kohonen map is trained, let's call it SOM_in. It will pre-process the input data. A second Kohonen map is trained called SOM_out. It is trained to classify the output examples. Learning perseptron in the input it has cluster numbers from SOM_in and the response is a cluster from SOM_out. And another perseptron is trained to "decode" the cluster number of SOM_out into the final prediction data. ??? I don't think anyone's tried that. It's a little tricky. But you try it. Tell me later... 1...31323334353637383940 New comment
so what are the mystery pictures?
there's nothing mysterious there...
so what's with the mystery pictures?
The currency indexes, there is nothing mysterious about them. )))
The indices are currency indices, there's nothing mysterious about them. )))
and what do they show?
and what do they show?
I have visions on Friday nights, too.
with an hors d'oeuvre, with an hors d'oeuvre!
I have visions on Friday nights, too.
with an hors d'oeuvre, with an hors d'oeuvre!
It's like looking at )))) The blue one is the euro, the green one is the quid. Eurobucks chart . I will not write a clever article. You can see it all.
What and why?
Hi there.
Has anyone tried the following type of construction:
- The first Kohonen map is trained, let's call it SOM_in. It will pre-process the input data.
- A second Kohonen map is trained called SOM_out. It is trained to classify the output examples.
- Learning perseptron in the input it has cluster numbers from SOM_in and the response is a cluster from SOM_out.
- And another perseptron is trained to "decode" the cluster number of SOM_out into the final prediction data.
???Hi there.
Has anyone tried the following type of construction:
- The first Kohonen map is trained, let's call it SOM_in. It will pre-process the input data.
- A second Kohonen map is trained called SOM_out. It is trained to classify the output examples.
- Learning perseptron in the input it has cluster numbers from SOM_in and the response is a cluster from SOM_out.
- And another perseptron is trained to "decode" the cluster number of SOM_out into the final prediction data.
???I don't think anyone's tried that. It's a little tricky. But you try it. Tell me later...