What are the machine learning algorithms, which ones are most commonly used, and what are their advantages?

The owner must not know much about machine learning to ask such a question. This question seems to be as specialized as the term "machine learning engineer". \x0d\\ machine learning, basic PCA modeling theory, Bayesian, boost, Adaboost, \x0d\ pattern recognition in a variety of features, such as Hog, Haar, SIFT, etc. \x0d\ deep learning in the DBN, CNN, BP, RBM, etc.. \x0d\\ Non-professional background, just a little bit of understanding. \x0d\\\\\x0d\\\ There is no commonly used, just for the needs of a specific design, or need to design a new suitable algorithm, now the most popular is considered CNN (convolutional neural networks) convolutional neural networks. In the case of a large enough sample, can get very accurate recognition results. Generally can 95% + correct rate. \x0d\ disadvantages: high hardware requirements, CUDA parallel framework is considered to be used very hot. But the general desktop computer to run a Demo spend a long time resource consumption is high. But this is also the general problem of this piece of algorithm.