Early diagnosis of lung cancer based on imbalance class data
Principal Investigator
Name
Xiaowen Yang
Degrees
postgraduate
Institution
Chongqing University of Posts and Telecommunications
Position Title
student
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-967
Initial CDAS Request Approval
Oct 18, 2022
Title
Early diagnosis of lung cancer based on imbalance class data
Summary
Medical data has typical class imbalances. The number of diseased samples is small relative to the number of healthy samples. This sample imbalance problem will make the model prediction results unsatisfactory, so it is important to solve the class imbalance problem when using medical data for corresponding studies.
The class imbalance data were balanced by resampling method, and high-quality samples were selected from the majority of classes, and balanced data sets were formed with a small number of samples and applied to the early auxiliary diagnosis of lung cancer.
The class imbalance data were balanced by resampling method, and high-quality samples were selected from the majority of classes, and balanced data sets were formed with a small number of samples and applied to the early auxiliary diagnosis of lung cancer.
Aims
1.Consider using an adaptive undersampling method based on step-by-step learning to efficiently select high-quality samples and balance the data distribution.
2.Use a cost-sensitive approach (Optimize Loss Function Penalty Terms) to make the model more focused on a small number of class samples.
3.Diagnose whether a patient has lung cancer-related disease based on early clinical imbalance data.
Collaborators
Chongqing University of Posts and Telecommunications