"New Neural" is an Open Source neural network engine project for the MetaTrader 5 platform. - page 18

[Andrey Dik]

In general, ideally, you should get something like this http://www.peltarion.com/products/synapse/comparison.php.

At least you can be guided by what inquisitive mind of a non-programmer user may want. :)

[TheXpert]

Mischek:

Ouch ! That's not what I know ?

Unless you can guess ... I think I've only told Victor so far.

Are you intriguing, bragging, inviting ? // underline as appropriate

per worker's request :)

I'll tell you right off the bat -- I don't think you'd be interested.

[TheXpert]

joo:

In general, ideally, it should be something like this http://www.peltarion.com/products/synapse/comparison.php

Mm-hmm. You're a killer. It's a 1000 grn license... You're looking up to me :)

[Andrey Dik]

TheXpert:
Uh-huh. You're a killer. The license costs 1,000 grivnas... You've got a good point :)

Compared to the capabilities of Neurosolutions and NSDT, you can say for nothing.

[Mykola Demko]

joo:

In general, ideally, you should get something like this http://www.peltarion.com/products/synapse/comparison.php.

At least you can be guided by what inquisitive mind of a non-programmer user may want. :)

Now explain me a lamer, which of these mats are neural network types, types of learning and what are trademarks:

Data input------------------

Text
Image
Signal generator
SQL database
User script
User custom built
Data output------------------
Text
SQL database
User custom built
Preprocessing---------------
Data rearrangement
PCA
Outlier removal
Query-based filtering
Statistical
User script
Interactive visualization
Postprocessing-------------
Error analysis and tracing
Confidece analysis
Sensitivity analysis
Real-time testing
Adaptive paradigms-------
Ststic neural networks
Dynamic neural networks
Kernel machines
Bayesian methods
LSTM
Fuzzy logic
Optimazation paradigms-




Genetic algorithms
Particle swarm
Monte carlo
Architectures---------------
MLP
Generalized MLP
Modular netwoks
Self-organizing maps
Neural gas
Competitive learning
Hebbian
FFCPA
Radial basis networks
LSTM
Time lagged recurrent
Partially recurrent
Wavelet networks
Fully recurrent
Neuro-fuzzy
Support Vector Machines
Custom architectures
System design---------------
Component based
Plug-in based xml-ph-NET compliant
User interface----------------
Graphical construction
Scalable user interface
Block embedded GUI
Deployment-------------------
Solutions are deployable
.NET compatible components
Reusable components
Scalable deployment
Royalty free use of deployed systems

[TheXpert]

Urain:

Now explain to me, as a lamer, which of these mate types of neural networks, types of learning and what are trademarks:

MLP
Generalized MLP
Modular netwoks -- ugh
Self-organising maps
Neural gas
Competitive learning -- unclear
Hebbian -- dunno, but if what I think is interesting
FFCPA -- dunno
Radial basis networks
LSTM -- ugh
Time lagged recurrent
Partially recurrent
Fully recurrent I wonder how one differs from the other.
Wavelet networks - I don't know
Neuro-fuzzy - this is probably what I was talking about.

Support Vector Machines - dunno

xz -- something I'm not familiar with

[Andrey Dik]

Learning/optimization algorithms:

Genetic algorithms
Particle swarm
Monte carlo

Network Types:

MLP
Generalized MLP
Modular netwoks
Self-organising maps
Neural gas
Competitive learning
Hebbian
FFCPA
Radial basis networks
LSTM
Time lagged recurrent
Partially recurrent
Wavelet networks
Fully recurrent
Neuro-fuzzy
Support Vector Machines
Custom architectures

[Andrey Dik]

We've already seen the possibilities of visual block programming in MQL5. So, everything is possible, it's up to the enthusiasts. :)

[Mykola Demko]

At least some kind of battle plan.

By the way, why don't they have BeckProp? Or did I misunderstand something.

There's only Genetics, Bee Swarm and Montecarlo?

[TheXpert]

joo:

By the way, just a side note. Most packages don't have genetics for networks at all. And in this list, it is only for checkboxes or if heuristics are missing (imho).