ANNs for improving Regression Analysis
- Simulations based on Linear Models
- Tools for the Appliances of ANNs to discover mediation and moderation effects
- Comparisons between classical statistical methods and the proposed ANNs
Creates one sample datasets with almost 8000 randomly drawn variables from multiple distributions. The simulation is model-agnostic, generated variables are all independent. All variables are named according to their distribution. The regeneration of the dataset can be triggered with the underlying function. Only input parameters are n and seed. By usining this routine, we can assure reprocability for all our results.