A Method of Random Forest Classification based on Fuzzy Comprehensive Evaluation

At present, the method of random forest classification is one of the most common classifier. However, when there are many attributes in the acquired data set and the weights of attributes are not the same or close to each other, the direct use of this method for data training will not only ignore the correlation between attributes, but also make it difficult to reflect the advantages of different attribute weights, resulting in a negative impact on the accuracy of the final results. In addition, if there are many attributes in the data set, in order to meet the normalization of attribute weights, the values of the weights will inevitably be small, which will also lead to the loss of information, resulting in a negative impact on the final training result of the random forest. In order to solve these problems, this paper proposes a random forest method based on fuzzy comprehensive evaluation, which can not only considers the correlation between attributes, but also preserves the information in the original data set to the greatest extent, and at the same time significantly improves the accuracy of the results of random forest training.