AI-Driven Forecasting of Cancer Trends: Integrating Machine Learning and Numerical Modeling

Abstract

In the current study, the incidence, mortality and survival rates of 12 different types of cancer are computed for the time period 2020-2024, based on numerical techniques like Runge-Kutta 4th Order (RK4) and machine learning methods, like Random Forests and Artificial Neural Networks. Future incidence, mortality and survival projections were performed for the years 2025 to 2050, which gave benefit information for public health planning for the future. The results indicate that cancer rates in the lung, liver, pancreatic cancer have increased while survival rates for other types of cancer such as thyroid and nonmelanoma skin cancer have improved due to early detection and treatment. Though some cancers, such as pancreatic and liver, continue to be a significant issue in that they do not have high survival rates. The ARIMA model's future predictions are in good agreement with the historical data, and are useful for forecasting the trends of cancer, but there are also some restrictions, such as the simplifying assumptions and lack of geographic conditions. The results highlight the importance of continuous research and targeted interventions in responding to high-risk cancers and enhancing treatment outcomes.

Key words; Cancer incidence, Mortality rates, Survival rates, Machine learning, ARIMA model, Early detection, High-risk cancers