Short Bio

Jeremy is a university assistant and PhD student in his third year at TU Wien. He is part of the research group Computational Statistics (CSTAT) and works under the supervision of Univ.-Prof. Peter Filzmoser. Jeremy’s research project focuses on Robust Statistics for functional data analysis (FDA). Within this project, he develops new methods for robust outlier detection and covariance estimation that can be helpful in various scenarios like functional principal component analysis (FPCA). Jeremy studied “Statistics and mathematical methods in economics” at TU Wien, earning him his BSc degree in 2020 and Dipl. Ing. degree in 2022. During his study, he focused on statistical methodology, culminating in a master thesis, “Dimension reduction for compositional data with weights based on graph theory.”

PhD Project - Robust Covariance Estimation and Outlier Detection for Functional Data

Supervised by Peter Filzmoser

In contrast to classical multivariate statistics, functional data contains infinite-dimensional observations. This leads to the challenge of adapting statistical methods, as common multivariate approaches cannot be applied directly to functional data. In the first stage of his research, Jeremy developed an efficient algorithm for outlier detection of functional data based on a robust covariance estimator. As part of the project Generalized relative data and robustness in Bayes spaces, his most recent work focused on relative functional data. Typical examples are age-specific fertility rates or spectra analysis of different materials. Currently, Jeremy is working on an adaptation of his initial algorithm to multivariate functional data. In these settings, each observation contains several variables, continuously observed over time.

Publications and Conferences

Papers and Proceedings

• Oguamalam, J., Filzmoser, P., Hron, K., Menafoglio, A., Radojicic, U. (2024). Robust functional PCA for density data. Submitted for publication. arXiv Preprint: https://arxiv.org/abs/2412.19004
• Oguamalam, J., Radojicic, U., Filzmoser, P. (2024). Functional Outlier detection. In: Ansari, J., et al. Combining, Modelling and Analyzing Imprecision, Randomness and Dependence. SMPS 2024. Advances in Intelligent Systems and Computing, vol 1458. Springer, Cham. https://doi.org/10.1007/978-3-031-65993-5_40
• Oguamalam, J., Radojicic, U., Filzmoser, P. (2024). Minimum regularized covariance trace estimator and outlier detection for functional data. Technometrics, 66(4), 588 − 599. https://doi.org/10.1080/00401706.2024.2336542

Presentations

• Oguamalam, J., Filzmoser, P., Hron, K., Menafoglio, A., Radojicic, U. (2025) Robust functional PCA for density data. 7th Joint Statistical Meeting of the Deutsche Arbeitsgemeinschaft Statistik (DAGStat), Humboldt Universität zu Berlin, Berlin, Germany.
• Oguamalam, J., Radojicic, U., Filzmoser, P. (2024). Robust covariance estimation and functional anomaly detection based on the minimum regularized covariance trace estimator. Austrian Statistical Days, TU Wien, Vienna, Austria.
• Oguamalam, J., Radojicic, U., Filzmoser, P. (2023). Functional outlier detection based on the regularized covariance trace estimator. Data Science, Statistics & Visualisation (DSSV), Universiteit Antwerpen, Antwerp, Belgium.
• Oguamalam, J., Radojicic, U., Filzmoser, P. (2023). Minimum regularized covariance trace estimator and outlier detection for functional data. Olomoucian Days of Applied Mathematics 2023 (ODAM), Palacky University Olomouc, Olomouc, Czech Republic.