Double-accuracy weighted random forest algorithm based on particle swarm optimization

Abstract

The invention discloses a double-accuracy weighted random forest algorithm based on particle swarm optimization, and the algorithm comprises the following steps: S1, determining an original data set D, randomly determining the number K of decision trees and the number m of features of the decision trees, and determining a pre-test sample rate X when m is smaller than or equal to n; s2, sampling the training data set Sk by using a Bootstrap sampling method to obtain an out-of-bag data Ok training sub Tk; s3, generating a kth decision tree according to a C4.5 algorithm, and calculating the finalweight of the decision tree; s4, repeating S2 and S3 until the number of decision trees is K; s5, using the decision tree set to test and divide the test data; and S6, taking the accuracy obtained inthe S5 as a fitness value of the particle swarm, performing iterative optimization by adopting a particle swarm optimization algorithm, comparing with historical classification accuracy, and finallyselecting an optimal model parameter. According to the invention, the accuracy of the random forest model is improved; Meanwhile, the problem that the algorithm accuracy is reduced due to parameter selection based on experience is avoided, and the performance of the algorithm is further improved.

Description

technical field [0001] The invention relates to a double-accuracy weighted random forest algorithm based on particle swarm optimization, belonging to the field of data processing. Background technique [0002] Random forest is a supervised ensemble learning classification technique. Its model is composed of a group of decision tree classifiers. The model classifies data by collectively voting on the classification results of a single decision tree to determine the final result. This average voting method may lead to a poorly grown decision tree affecting the final classification result, and it is prone to "tie votes". The traditional random forest algorithm fully guarantees the independence and difference between each decision tree by injecting randomness into the training sample space and attribute space, and overcomes the over-fitting problem of the decision tree well. It also has better robustness. However, due to the randomness of training samples and attributes and th...

AI Technical Summary

Problems solved by technology

However, due to the randomness of training samples and attributes and the uncertainty of decision tree trees, the actual effect is not ideal.

Images