"I'm currently working with multivariate data containing 18 features, and I'm looking to understand the underlying distribution of these features. My goal is to leverage this information to enhance the performance of a neural network for classification tasks. The 'Pearson Family of Distribution' is commonly used in univariate data to identify the underlying distribution. Is there an equivalent method or approach, like 'Pearson Curves,' that can help determine the underlying distribution of multivariate data?
Any insights or guidance on this matter would be greatly appreciated.