Saturday, December 23, 2017

Nebula start editon for MT4

Nebula start editon for MT4


Advisor to the two moving averages. The advisor is implemented self-learning mode, which allows you to immediately begin to trade, but also can conduct a preliminary training advisor for more efficient trade. Default EA is set up to work on timeframe H1.


The basis of the adviser is modified probabilistic neural network. Initially the neural network does not contain neurons, in this mode, the adviser does not produce trade, and only trained. In operation, the advisor is self-learning and neural network neurons are responsible for one of two classes - buy or sell.


As soon as there is at least one neuron, the adviser can trade.




settings



  • Target type - target of type:

    • Points - in points;

    • Money - money;



  • TargetProfit, TargetLoss, TrailingStop - setting target levels of income, loss and trailing stop level in points. It applies to the symbol as a whole;

  • VolumeLot - the volume of trading lot;

  • Slippage - level slippage in points;

  • WaitPeriod - pause after the transaction, in milliseconds;

  • TradeAttempt - the number of trading attempts to enter the market;

  • MAFastPeriod - a period of rapid moving average;

  • MASlowPeriod - period of slow moving average;

  • MAMethod - mode of calculation of moving averages;

  • MAAppliedPrice - type of prices for the calculation of moving averages;

  • NNActivationLevel - the level of activation of neurons. This is the minimum value of the output network that affects the decision. Recommended value - not less than 0.95;

  • NNCalcTime - interval computation neural network in minutes;

  • EnableNNload - neural network to download data from a file;

  • FilePrefix - setting a file name to store the data of the neural network. The file is created in the common terminal folder in Files \ NN2MA;

  • EnableInfoPanel - display the information panel, where displays information about the neural network and sales levels;

  • Expert Magic number - the unique identifier of the expert;

  • Order comment - Comment to the order.




Algorithm prior training or adjustment counselor




  1. Set the zero value TargetProfit parameters, TargetLoss and TrailingStop. EnableNNLoad values ​​and EnableInfoPanel - No. FilePrefix Specify the file name.


    Set values ​​optimize the two parameters of moving averages and MAFast MASlowPeriod from 1 to 100 or more.


    Select the period of optimization and testing period, for example, the optimization period in 2016, the test period - 2017 Start optimization on the period in 2016.




  2. After the completion of the optimization, select the appropriate results. In the selection of the results, follow these guidelines: MAFastPeriod value must be less MASlowPeriod, other indicators were selected on the discretion.


    Run a single test counselor with the selected parameters in the period 2016. If the figures obtained will suit you, you can check the adviser on the test period in 2017.


    As a result of a single test file with the parameters of the neural network will be created, which can be used for a counselor in real trading mode. With each new start testing with a single neural network configuration file is overwritten.




  3. Run a single test on a test period of 2017. When this parameter is set EnableNNLoad - Yes, it will load the configuration of the neural network, resulting from single testing period in 2016. If the test results you are satisfied, then the adviser is ready to trade. TargetProfit parameters, TargetLoss TrailingStop and can be set to your liking.




Nebula start editon for MT4

No comments:

Post a Comment