Auto Encoder Based Feature Learning and Up-sampling to Enhance Cancer Prediction

  • M. Akkalakshmi, Y. Md. Riyazuddin, V. Revathi, Abhisek Pal


Early detection of cancer is vital in long term survival of the patient.  Cancer detection requires skilled doctors and analysis of patients data in different form like images, clinical data and gene expressions. AI can help in analysing large and complex data of different forms and still achieve good accuracy on par with specialist doctors. Semi-supervised learning techniques of AI can deal with even scarce and incomplete datasets. This paper deals with up-sampling the unbalanced breast cancer datasets and   proposes a semi-supervised learning approach for latent features learning using autoencoders to improve the prediction accuracy which helps in cancer diagnosis and treatment.