A Heuristic Machine Learning-Based Optimization Technique to Predict Lung Cancer Patient Survival.

Area of Science:

• Oncology
• Bioinformatics
• Machine Learning

Background:

• Lung cancer poses a significant global health challenge with high mortality rates.
• Accurate prediction of patient survival is crucial for effective cancer management.
• Existing models often rely on single data types, limiting predictive power.

Purpose of the Study:

• To develop an accurate model for predicting overall survival time in lung cancer patients.
• To address the limitations of current models by integrating multiple data sources.
• To improve the precision of survival predictions beyond a simple five-year outlook.

Main Methods:

• A novel approach combining Naive Bayes and SSA ( [correction: SSA is not defined in the abstract, assuming it's a typo or needs clarification] ) was employed.
• Two machine learning tasks were formulated: binary classification for five-year survival and regression analysis for survival time estimation.
• The model was evaluated using metrics such as accuracy, recall, precision, and mean absolute error.

Main Results:

• The proposed Naive Bayes and SSA technique achieved high performance metrics: 98.78% accuracy, 98.4% recall, and 98.6% precision.
• The model demonstrated a mean absolute error of prediction within one month for overall survival time.
• Biomarker genes associated with lung cancer were identified, contributing to model's predictive capability.

Conclusions:

• The Naive Bayes and SSA approach offers a significant advancement in predicting lung cancer patient survival.
• This method provides a more personalized and accurate estimation of survival time, aiding clinical practice.
• Further research can explore integrating more diverse data types to enhance predictive accuracy.