Modelling survival time distributions of cancer data using artificial neural networks

  • Rashmi Raj Joshi

Student thesis: Doctoral ThesisDoctor of Philosophy


The aim of the project is to develop and assess the effectiveness and usefulness of a novel approach for modelling time-to-event data (survival data) and survival time distributions, in comparison with the most popular conventional statistical approaches. The analysis of such data is termed survival analysis. Survival methods are used in many circumstances in medical and diverse areas of research. The research is in collaboration with the West Midlands Cancer Intelligence Unit (WMCIU), who supplied the malignant melanoma data set containing time-to-event data, upon which the analysis presented herein is based. Owing to the existence of a national population based cancer registration system, patients diagnosed as having cancer can now be tracked through their various treatment options to disease outcome and death. In addition to cancer incidence data, corresponding clinical and pathological factors are recorded at times of diagnosis and treatment. The dependence of disease outcome on these factors and covariates can therefore be assessed
Date of Award2004
Original languageEnglish
Awarding Institution
  • Coventry University
SupervisorColin Reeves (Supervisor) & Charles Johnston (Supervisor)


  • Survival Data
  • (WMCIU)
  • Cancer
  • makignant Melanoma
  • Survival Function
  • Hazard Function
  • Cox's PH Model
  • Survival Analysis

Cite this