Applies to: NeuralTools 5.x and Up

 

Predicting Multiple Dependent Variables in NeuralTools

You cannot train a single NeuralTools net with more than one dependent variable at a time. Each training session handles only one dependent variable, even if your data set contains multiple dependent variables.

How to work with multiple dependent variables:

1. Single data set approach:

   • Set up your independent and dependent variables in one data set.

   • Train a first net using the independent variables and the first dependent variable.

   • Train a second net using the same independent variables and the second dependent variable.

   • This produces two Live Prediction cells in the same data set, each reflecting predictions from its own net.

   • Between training sessions, you must change which variable is defined as dependent in the Data Set Manager; you cannot have more than one dependent variable defined at the same time.

2. Multiple data sets approach:

   • Train each net using a separate data set (one net per dependent variable).

   • Use both nets for Live Prediction in a third data set.

Key point: Multiple dependent variables are supported in your overall workflow, but each neural net handles only one dependent variable at a time.