Applies to: NeuralTools 6.x and Up

 

Combining Multiple NeuralTools Models (Ensemble Methods)

If you have two separate data sets with the same dependent variable and have trained them independently, you might wonder if you can combine them to get better predictions. This is an area known in machine learning as ensemble methods.

• Ensemble methods involve combining predictions from multiple models rather than selecting the single best model.

• The idea is that different models may capture different patterns in the data, and combining them can reduce variance, bias, or overfitting, leading to better overall predictive performance.

• NeuralTools itself doesn’t have a built-in “ensemble” feature for automatically merging trained networks. However, you could manually implement an ensemble approach:

  • Train separate nets on each data set.

  • For a given input, compute predictions from both nets.

  • Combine the predictions (e.g., by averaging, weighted averaging, or majority vote for classification).

References and further reading:

• Fensterstock, A., Salters, J., & Willging, R. (2013). On the Use of Ensemble Models for Credit Evaluation. The Credit and Financial Management Review, Q4.

• Dietterich, T.G. (2000). Ensemble Methods in Machine Learning.

• Kuncheva, L., & Whitaker, C. (2003). Measures of Diversity in Classifier Ensembles.

• Mitchell, T.M. (1997). Machine Learning.

• Schwenk, H., & Bengio, Y. (2000). Neural Computation.

• Sollich, P., & Krogh, A. (1996). Learning with Ensembles: How Overfitting Can Be Useful.

• Specht, D.F. (1989). Probabilistic Neural Networks.

• Zenko, B. (2004). Is Combining Classifiers Better Than Selecting the Best One?

Key takeaway: Combining models can improve prediction accuracy, but you will need to implement the combination yourself in Excel or using NeuralTools XDK.