Henrique Aguiar

Henrique completed undergraduate studies in Mathematics at the University of Oxford, graduating with a distinction in 2019. His Master’s thesis was on Deep Learning and studying how differential equations could be used to construct new Neural Network architectures. His interests span a wide range of areas in Statistics and Computer Science; in particular, its applications to healthcare. This includes topics such as Time-Series Analysis, Clustering and Generative Modelling. Henrique is currently working on Clustering Methods for time series electronic health records data.

Henrique's DPhil focuses on the development of clustering methodologies applied on temporal, multivariate, and multimodal Electronic Health Records (EHR) data (including vital-sign observations, laboratory test results, medications, blood tests, cohort and demographic data, etc.) that are capable of improving healthcare delivery. Some previous projects have included tackling severe imbalance in multiclass settings, development of a cluster-specific feature-time clinically interpretable mechanisms, and extending these to model patient evolution throughout admission.

I am currently working under the supervision of Dr Tingting Zhu on the use of vital-sign features, as well as cohort and outcome information, from hospital databases and public datasets, to improve prediction of patient trajectories and deliver personal monitoring recommendations to boost patient well-being.

H. Aguiar, M. Santos, P. Watkinson, T. Zhu (2020) Phenotyping Clusters of Patient Trajectories suffering from Chronic Complex Disease,

Aguiar, H., Santos, M., Watkinson, P., and Zhu, T. (2022). “Learning of cluster-based feature importance for electronic health record time-series”, in International Conference on Machine Learning.

Aguiar, H., Santos, M., Watkinson, P., and Zhu, T. (2020). “Phenotyping Clusters of Patient Trajectories suffering from Chronic Complex Disease”, in NeurIPS Workshop on Machine Learning for Healthcare.