Defining hyperparameters

Let's imagine that I want to optimize the period of two EMA indicators in my strategy.

This is how I would have defined them so far, by hard-coding periods as integer values; in this case 50 and 200:

@property
def slow_sma(self):
    return ta.sma(self.candles, 200)

@property
def fast_sma(self):
    return ta.sma(self.candles, 50)

To define the hyperparameters of your strategy, simply add the hyperparameters() method to your strategy. It must return a list of python dictionary objects; each of which has specific key values depending on the parameter type. Here's the code for this example:

def hyperparameters(self):
    return [
        {'name': 'slow_sma_period', 'type': int, 'min': 150, 'max': 210, 'default': 200},
        {'name': 'fast_sma_period', 'type': int, 'min': 20, 'max': 100, 'default': 50},
    ]

Parameter Types

Jesse supports the following parameter types:

Integer Parameters

For integer values, use 'type': int or 'type': 'int'. Required keys:

• name: Parameter name
• min: Minimum value (inclusive)
• max: Maximum value (inclusive)
• default: Default value
• step (optional): Step size between values (default: 1)

{'name': 'rsi_period', 'type': int, 'min': 5, 'max': 30, 'default': 14, 'step': 1}

Float Parameters

For decimal values, use 'type': float or 'type': 'float'. Required keys:

• name: Parameter name
• min: Minimum value (inclusive)
• max: Maximum value (inclusive)
• default: Default value
• step (optional): Step size between values (default: None, which allows any float value in the range)

{'name': 'entry_threshold', 'type': float, 'min': 0.1, 'max': 5.0, 'default': 1.0, 'step': 0.1}

Categorical Parameters

For parameters that should be selected from a list of options, use 'type': 'categorical'. Required keys:

• name: Parameter name
• options: List of possible values (can be any type)
• default: Default value (must be one of the options)

{'name': 'indicator_type', 'type': 'categorical', 'options': ['rsi', 'macd', 'bollinger'], 'default': 'rsi'}

Boolean Parameters

For boolean parameters, you can use either:

Method 1: Use categorical type

{'name': 'use_filter', 'type': 'categorical', 'options': [True, False], 'default': False}

Method 2: Use integer type with min=0, max=1

{'name': 'use_filter', 'type': int, 'min': 0, 'max': 1, 'default': 0}

Then in your code, convert the value to boolean: bool(self.hp['use_filter'])

Accessing Hyperparameter Values

Jesse injects each hyperparameter into the self.hp property that is available in your strategy class.

Now let's rewrite the starting example to use the dynamic hyperparameters:

@property
def slow_sma(self):
    return ta.sma(self.candles, self.hp['slow_sma_period'])

@property
def fast_sma(self):
    return ta.sma(self.candles, self.hp['fast_sma_period'])

If I execute a backtest, I will get the same results as before as if I was using the hard-coded values. That's because I defined 50 and 200 as default values for my hyperparameters.

When running optimization, Jesse will search through the parameter space defined by the min, max, and step values (for numeric parameters) or through the available options (for categorical parameters) to find the optimal combination.