Posterior probability algorithm based on weighted decision tree

Abstract

The invention discloses a posterior probability algorithm based on a weighted decision tree. According to the method, based on a conventional decision tree algorithm, a weight influence factor is added when a decision tree is split, and the influence of signal-to-noise ratio classification is improved, so the final information entropy of nodes with high signal-to-noise ratio is enhanced, the overall probability distribution relation is effectively changed by the changed information entropy, and the influence of untrusted classification is reduced. According to the invention, the error influence caused by abnormal data can be effectively reduced; the dual polarization phenomenon caused by unbalanced data sample distribution is reduced; the influence of credibility is brought in, so the reliability of the model is improved; and an over-fitting phenomenon caused by data exception is reduced.

Description

technical field [0001] The invention belongs to the posterior probability algorithm based on weighted decision tree, relates to the field of data structure, and is especially applicable to the method in the field of data mining Background technique [0002] A decision tree is the process of classifying data through a set of rules. It provides a rule-like way of getting what value under what conditions. Decision trees are divided into classification trees and regression trees. The classification tree is a decision tree for discrete variables, and the regression tree is a decision tree for continuous variables. [0003] Recent surveys show that decision tree is also the most frequently used data mining algorithm, and its concept is very simple. A very important reason why the decision tree algorithm is so popular is that users basically do not need to understand the machine learning algorithm, nor do they need to delve into how it works. Intuitively, a decision tree classif...

Images