Classification of Cancer Subtypes Based on Multi-Granularity Cascade Forest

EasyChair Preprint no. 8421

Abstract

Treatment options are different for different cancer subtypes.It is of great significance for cancer patients and medical field to determine the types of cancer subtypes in time.For some redundant genes and noisy genes in the sample data of cancer subtypes,decision trees were used for feature selection,which effectively improved the classification performance of the classification model.In order to solve the problem of over-fitting caused by traditional machine learning methods in classification of cancer subtypes,multi-granularity Cascade forest (gcForest),an algorithm combined with machine learning and deep neural network,was applied.Comparing gcForest with support vector machine,logistic regression,random forest and K-nearest Neighbor method,the experimental results show that gcForest has better performance than other traditional machine learning algorithms.

Keyphrases: cancer subtype data, Decision Tree, gcForest, multi-classification

@Booklet{EasyChair:8421,
  author = {Chunxiao Jiang and Hua Duan},
  title = {Classification of Cancer Subtypes Based on Multi-Granularity Cascade Forest},
  howpublished = {EasyChair Preprint no. 8421},

  year = {EasyChair, 2022}}