I found a slide where it says that the ways of studying in computer systems are divided into:

  1. Deductive, or analytical, training (expert systems). There is knowledge, formulated by an expert and somehow formalized. The program derives specific facts from these rules and new rules.
  2. Inductive learning (statistical learning). Based on empirical data, the program builds a general rule. Empirical data can be obtained by the program itself in previous sessions of its work or simply presented to it.
  3. Combined training.

Please provide the types and algorithms of combined learning.

  • apparently found the source of the information, this is a popular science lecture N. Yu. Gold [How are machines trained?] [1] and there it is more precisely stated: - Combined training, containing elements of both deductive and analytical training. [1]: uic.unn.ru/~zny/ml/Others/ml_pop.pdf - Merlin

1 answer 1

I found a nice description. And it (it seems) is made with the help of a manin redo. But you can see the original. Selection of Microsoft Azure Machine Learning Algorithms Memo on machine learning algorithms https://docs.microsoft.com/ru-ru/azure/machine-learning/studio/algorithm-choice 04/25/2017 Read Time: 29 min

Briefly, with illustrations of graphs. Best.