Structured data vs. unstructured data problems

[mamurak]

In the current exercise a neural network is trained on the Diabetes data set. To educate the students on which general use cases exist and which training frameworks would typically be used, I propose to modify this chapter as follows:

• Following the discussion of the machine learning concepts, introduce the concepts of structured and unstructured data with practical examples.
• Give a short overview of the most common training frameworks and typical use cases with respect to structured and unstructured data problems, e.g. scikit-learn, xgboost, lightgbm, catboost, tensorflow, pytorch.
• Change the model trained in the existing Diabetes problem to a scikit-learn type model.
• Add another exercise for working on an unstructured data problem, for which a deep learning model with PyTorch is used.

[noelo]

+1