Almost all models for machine learning is statistics based, expert system is a bit different as it’s based more on known logics.
Biology-based
Neuralnet
Neuralnet (Artificial Neural Network) or ANN is a biology-inspired model for machine learning. Neuralnet is the most common ML model of the modern time; possibly later we may have biology-simulated network instead. This book will use mostly only neuralnet for use in machine learning.
Deep Learning is neuralnet-based learning, either on software or GPU.
Some Abbreviations
ANN — Artificial Neural Network
CNN — Convolutional Neural Network
Known-logic-based
Expert System
Expert system is the learning based a lot on known logics, almost all if-else. Deep Blue was one of those.
Mathematics-based
SVM
SVM (Support Vector Machine) is a mathematics-based model for machine learning which utilise minimal perpendicular vectors for minimisation of distance. SVM is rather limited in solving problems, it is binary classifier. It can not do regression and in order to do multiple classifications, SVM must work in the one-vs-all strategy again and again.
