Lung cancer risk prediction method based on feature selection and artificial neural network
Nan-Nan Xie1, Liang Hu, Tai-Hui Li
1Computer Science and Technology College, Jilin University, Changchun, China
Asian Pacific Journal of Cancer Prevention : APJCP
|January 6, 2015
Summary
This study introduces a new lung cancer risk prediction method using Fisher and ReliefF feature selection with BP Neural Networks. The approach accurately identifies risk factors for health monitoring and self-testing.
Area of Science:
- Oncology
- Biostatistics
- Machine Learning
Background:
- Lung cancer remains a leading cause of cancer-related mortality worldwide.
- Accurate risk prediction is crucial for early detection and intervention.
- Current prediction models may involve numerous risk factors, complicating practical application.
Purpose of the Study:
- To develop an efficient and accurate method for predicting lung cancer risk.
- To reduce the number of risk factors required for prediction without sacrificing accuracy.
- To create a framework suitable for health monitoring and self-testing applications.
Main Methods:
- A two-step approach combining Fisher and ReliefF feature selection algorithms.
- Utilizing BP Neural Networks for risk prediction based on selected factors.
- Development of a specific algorithm named LCRP (lung cancer risk prediction).
Main Results:
- The combined feature selection effectively identified an appropriate quantity of risk factors.
- The LCRP algorithm demonstrated satisfactory prediction accuracy.
- The method achieved high accuracy even with a reduced set of risk factors (low dimensions).
Conclusions:
- The proposed method offers a viable solution for lung cancer risk prediction.
- The LCRP algorithm is practical for health monitoring and self-testing due to its efficiency.
- This approach shows promise for improving early lung cancer detection strategies.
