Analysis methods commonly used in 5

One, the line chart analysis method and the hash chart analysis method

Second, covariance (two parameters) and covariance matrix (parameters greater than two)

Three, correlation coefficient method (can directly use Excel to find correlation coefficient)

Four or one-yuan regression (two parameters) and multivariate regression (extra two parameters)

V. Information entropy and mutual information

Mutual information refers to the degree of association between two random variables, that is, given a random variable, the degree of uncertainty of another random variable, so that the mutual information value of the minimum of 0, meaning that given a random variable is not related to the determination of one other random variable, the maximum value is the entropy of a random variable, which means that given a random variable , which can completely eliminate the uncertainty of another random variable. Popular speaking is:

So I'm not sure about X (the uncertainty is H (X)), and I'm telling me y I'm h for x uncertainty (x| Y), this reduction in uncertainty is mutual information I (X; Y) =h (X)-H (x| Y

Each of these methods has its own characteristics. The graph method is the most intuitive, the correlation coefficient method can see the correlation between variables 22, the regression equation can be used to refine the correlation relation, and generate the model for prediction, mutual information can measure the correlation relation between the text class features.

Analysis methods commonly used in 5