A Failed Design for Dynamic Multiple Logistic Regression in MQL5

Summary

The author describes an unfinished attempt to build a multiple logistic regression library in MQL5 that can handle a variable number of independent data columns. Earlier functions were hard-coded for particular numbers of predictors, which the author considers repetitive and difficult to extend. The proposed interface instead accepts a string listing the predictor columns, then aims to apply the same calculations across them.

Because the author sees no practical way to create the needed arrays dynamically, they hypothesize storing all columns in one array and copying each column's rows into a working array during iteration. They also consider dynamically created CSV files but worry about file handling overhead and slower loops. The attempt remains explicitly unsuccessful: the post presents the array layout as a hypothesis and invites suggestions, without a completed algorithm, validation, or predictive results. It is useful as an implementation problem discussion, not as a tested regression method.

Key ideas

• The design goal is to support a variable number of predictors without writing separate functions for each model size.
• The author proposes identifying predictor columns through a string parameter.
• The proposed storage approach places column data in one array and copies each column for processing.
• The author considers dynamic CSV files but worries that repeated file access may be costly.
• The approach is unfinished and provides no validated model or predictive evidence.

Tags

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