Classification Based on Nature of Inputs:

Classification:

In such an approach, each of the inputs is given a weightage parameter depending on their influence on the output. So, our prediction is based only on the values of the future inputs.

Non-probabilistic Models:

Nature of the problem (Regression, Classification, Clustering, etc.) – Number of data records available for training – Decision support or cognitive use case? – Need for explain-ability

Classification Based on the Solution:

Classical Machine Learning:

Decision trees and nearest neighbors algorithms are examples of this category.

1.3 Classification of Machine Learning Algorithms