School of Science, Engineering and Information Technology

Using deep learning methods to predict anomalies for healthcare applications

Project Title:

Using Deep Learning Methods to Predict Anomalies for Healthcare Applications


  • Associate Professor Shyh Wei Teng (Principal)
  • Dr Jiangang Ma
  • Professor Guojun Lu
  • Professor Manzur Murshed

Contact person and email address:

  • Dr Jiangang Ma:

A brief description of the project:

Currently the use of a large number of mobile devices, internet of things and sensors in healthcare applications has produced a huge amount of data. Especially in the field of healthcare, the general use of portable signs sensors (e.g. mobile phones, hand ring) makes it possible to collect the patient’s signs signals (e.g., saturation of peripheral oxygen, end-tidal oxygen and end-tidal carbon dioxide) at any place and at any time. For example, monitoring patient in intensive care units (ICU) is a critical but complex task because it requires the analysis of a tremendous amount of streaming data collected from many sensors and in a timely fashion. However, existing approaches of predicting anomalies have some limitation such as inefficient and not fully automatic.

This project's goal is to develop a novel deep learning method for predicting anomalies from data streams for healthcare applications. This includes the development of quantitative models, algorithms and software tools for patients that can be used to predict health status, as well as to help prevent disease or disability through using deep learning methods. In this project, you will develop novel unsupervised deep feature learning method to derive a general-purpose patient representation from data, which facilitates clinical predictive modelling. You will develop original algorithmic and model solution in the time series motif discovery to reducing the computational complexity and avoiding unexpected repetitions among different motifs and cloud-based platform (cloud storage and cloud computing) as a carrier data processing.