Using Discriminant Analysis to Classify Market Direction

Summary

The article explains how discriminant analysis can classify market observations into groups such as upward and downward price moves. It illustrates the idea with technical indicators: statistical analysis can identify variables that contribute little to separating the groups, then produce a scoring equation for each group. At prediction time, the indicator values are inserted into the equations and the larger score determines the forecast class. The workflow covers collecting indicator and price data in an Expert Advisor, preparing labeled observations, selecting variables, evaluating a model with test data, and translating the resulting equations into a trading system.

A worked FOREX example reports a classification accuracy of about 55% and supplies discriminant equations based on selected indicator variables. The author notes that the indicators and analysis period were chosen somewhat arbitrarily, so the example demonstrates a process rather than a validated trading edge. Results depend on the sample and model choices; the tutorial does not establish robustness across markets or future data.

Key ideas

• Discriminant analysis uses predictor variables to assign observations to predefined outcome groups.
• The example labels data according to the direction of a later price bar and records indicator values as predictors.
• Variable selection can remove indicators that contribute little to separating the groups.
• Group-specific scoring equations classify new observations by comparing their scores.
• The example's reported accuracy is modest, and its arbitrary settings limit claims about predictive value.

Tags

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