View source for The healthcare data mining with advance AI technology

{{StudentProjectTemplate
|Summary=The cardiovascular health care project
|Keywords=Time series prediction and imputation, classification
|TimeFrame=Fall 2024
|References=[1] Review of multimodal machine learning approaches in healthcare, Information Fusion, Vol 114, February 2025. 
[2] Predicting Chronic Kidney Disease using a multimodal Machine Learning approach, Aakruti Mishra and Navaneeth Puthiyandi, Stockholm University.
|Supervisor=Guojun Liang, Prayag
|Level=Master
|Status=Open
}}
Our project aims to analyze complex healthcare data through the innovative application of advanced AI technologies, such as diffusion models machine learning frameworks. By leveraging these techniques, we seek to uncover latent structures within time series data, addressing challenges like causal discovery, missing data imputation, and accurate future prediction—ultimately enhancing clinical decision-making and patient outcomes.

Healthcare data is inherently diverse, encompassing clinical time series (e.g., vital signs, lab results) and medical imaging (e.g., chest X-rays). Traditional methods often overlook the relationships between variables such as heart rate, blood pressure, and respiratory rate, which are crucial for understanding patient health. Our project proposes the development of a machine learning model using the MIMIC dataset, integrating both clinical time series and medical imaging data to predict outcomes like patient mortality, disease progression, and treatment response (e.g., effectiveness of statin therapy).

The model will employ convolutional neural networks (CNNs) for feature extraction from imaging data, while diffusion models and transformers will process the time series data. The research will focus on fusion strategies that effectively combine temporal and image-based information, optimizing predictive accuracy and clinical interpretability.

Key topics include:

Time Series Prediction for Patient Health Monitoring:
Applying AI techniques to predict disease progression and detect health deterioration early, enabling proactive interventions in chronic condition management.

Time Series Imputation for Healthcare Data:
Using AI to reconstruct missing values due to incomplete records or sensor failures, supporting more reliable diagnostic and monitoring systems.

Causal Discovery in Healthcare Time Series:
Uncovering causal relationships within patient data to provide deeper insights into how medical variables interact, enhancing personalized treatment strategies.

By integrating multiple data sources and leveraging advanced AI techniques, this project aims to create a comprehensive framework for healthcare analysis, supporting better-informed clinical decisions and improved patient outcomes.