Predicting Thyroid Cancer Recurrence

Aims of the project

Things to explore when moving into model building

• Which will be the most performant features?
• Do we need all the features or will models perform better with fewer features?
• Which feature selection method will result in the best model performance?
• What impact does oversampling vs no oversampling vs Smote have on model performance?
  • In particular, will this result in being able to minimise false negatives (type II errors)?

References

• Cancer Staging - NCI [WWW Document], 2015. URL https://www.cancer.gov/about-cancer/diagnosis-staging/staging (accessed 7.20.24).
• Borzooei, S., Briganti, G., Golparian, M. et al. Machine learning for risk stratification of thyroid cancer patients: a 15-year cohort study. Eur Arch Otorhinolaryngol (2023). https://doi.org/10.1007/s00405-023-08299-w
• Welch Dinauer, C.A., Michael Tuttle, , Robie, D.K., McClellan, D.R., Svec, R.L., Adair, C. and Francis, G.L. (1998), Clinical features associated with metastasis and recurrence of differentiated thyroid cancer in children, adolescents and young adults. Clinical Endocrinology, 49: 619-628. https://doi.org/10.1046/j.1365-2265.1998.00584.x