Optimizing High-Frequency Indicator Weights for Trading Signals

Summary

The document describes a stock trading signal that combines multiple predictors with different forecast horizons. At each tick, the weighted indicator total is compared with a positive and negative threshold to determine whether to buy or sell. Differential Evolution is used to optimize the weights and threshold against an objective based on price differences between consecutive trades.

The central issue is computational cost on a very large tick dataset with many indicators. The document asks whether a faster method could solve the optimization problem, but gives no proposed alternative, performance comparison, or out-of-sample evidence. The stated objective also leaves execution costs, position handling, and risk constraints unspecified, so the optimization setup alone does not establish that the resulting signals would be profitable in live trading.

Key ideas

• The signal combines predictors with different horizons using a weighted sum.
• Buy and sell decisions occur when the combined score crosses symmetric thresholds.
• Differential Evolution optimizes both predictor weights and the threshold against a price-difference objective.
• The dataset size makes the proposed search computationally demanding.
• The document raises the need for faster optimization but does not evaluate an alternative.

Tags

# Fitting High Frequency Indicators


# Fitting High Frequency Indicators












I have a high frequency time series of the bid and ask prices of a stock recorded on every tick. For each data point I also have a certain indicators that predict the future movement of the price. The indicators have different horizons of the predictions, some being optimal at few second intervals and others few minutes. I need to assign these predictors weights and based on weather the linear combination crosses a threshold, the decision will be taken to buy of sell the stock. So far I have tried the Differential Evolution (DE) method to figure out the weights. I use a black box model with the weights vector($w_i$) and threshold as inputs. For each data point I have a vector of indicators($\alpha _i$). $$ total\_alpha = \sum\alpha _i*w_i $$ If $$ total\_alpha > threshold, BUY $$ Else If $$ total\_alpha < -threshold, SELL $$ The output of the model is the sum of difference between each between the price of each consecutive buy and sell. This output is being optimised by the DE algorithm. The issues with it being the computational aspects. I have large data sets of sizes(~7e8x20) and the time it takes for the DE algorithm. Is there a better and a faster way to solve this problem?

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.