Logistic regression algorithm introduced

# logistic regression algorithm (Logistic Regression)
1, according to 2017-- 2018 statistics and science in the field of artificial intelligence big data, different machine learning algorithm accounted ranked as follows, in which logistic regression, decision trees, random forest and artificial neural network algorithm accounted for the top four, the most widely used, followed by a Bayesian algorithm, integrated learning and the SVM algorithm. Currently the world's hottest cutting-edge artificial intelligence research RNN and CNN depth learning algorithm ranked ninth and tenth, mainly because of its high demand for data and requires high computational power.

2, logistic regression algorithm is the most widely used algorithm, although regression algorithms, but it solves a classification problem, rather than regression, it is the principle feature is the probability of occurrence of the samples in the sample, and the probability is a number, so it is called regression algorithm.

3, logistic regression algorithm which can be seen as a regression algorithm, can be seen as a classification algorithm, commonly used to solve classification problems, mainly binary classification problem , the problem is not suitable for multi-classification, can be deformed by a certain skill indirect resolved.

4, since the logistic regression for prediction result obtained is a probability of an event, it is a prediction value range between 0 and 1, and the probability that the conversion function is generally used in the sigmoid function, which converts the real number in the range 0 -1, and 0 is a boundary point, when t> 0, the probability p is greater than 0.5, when t <0, the probability p <0.5.


Guess you like

Origin www.cnblogs.com/Yanjy-OnlyOne/p/11347000.html