Signal And Scoring

Ensemble

Definition: Creates a predictive time series for a target asset based on selection of the top scoring inputs.

Syntax: ensemble(asset, method, daysLookBack, useTopScoreCount)

Arguments:

• asset: the symbol to forecast

• method: scoring method to use

• daysLookBack: the number of previous days to be used in scoring the input rows (default 30)

• useTopScoreCount: the number or percentage of top-ranked rows to retain in the ensemble of rows (default 20 percent)

Return: Single trendline time series derived from inputs

Description: Input to the action is a list of candidate time series. Its size must be at least as great as useTopScoreCount, if useTopScoreCount is a number rather than a percentage.

Example:

tl = inputs -> ensemble(asset=SPY, "pl", daysLookBack=3, useTopScoreCount=5)

Score Signal

Definition: This function implements the scoring methods used in Create Trendline, though it does not function in exactly the way that function uses them.

Syntax:

score_signal(target, method, closeonneutral, zerosize)

Arguments:

• target: time series for the symbol to forecast

• method: scoring method to use

• closeonneutral: whether to close positions when the signal enters the neutral zone

• zerosize: defines the bounds of the neutral zone (default 0.0)

Return: Time series of accumulated scores

Description: Input to the action is a trendline indicating long or short positions for trading.

Example:

spy_asset = SPY.close 
scores = tl -> score_signal(spy_asset, "pl")