My methodological research interests are Semiparametric efficiency, Distance-based modeling, Causal (mediation) inference, Missing data, and Psychometrics. I am delighted to apply the developed approaches to high-dimensional data burgeoning from different disciplines such as omics and wearable data, as well as the functional connectivity of brain fMRI data.
A Distance-Based Semiparametric Regression Framework for Dimensiona-Reduced Between-Subject Attributes
Motivated by the real data with astronomical dimensions, I am profoundly passionate about finding effective dimension-reduction metrics through “between-subject attributes” at the pairwise level, to differentiate from their classical “within-subject” counterparts that concern only one individual. Modeling such between-subject attributes designates between- and within-subject variability above and beyond the mean responses. Such variabilities are of interest in a growing number of studies. Adopting an efficient semiparametric framework yields robust inferences that allow for minimum model assumption but sensitive signal detections.
References:
Liu, J, Zhang, X, Lin, T, Chen, R, Zhong, Y, Chen, T, Wu, T, Liu, C, Huang, A, Nguyen, T, Lee, E, Jeste, D, Tu, XM. (2023). A new paradigm for high-dimensional data: distance-based semiparametric feature aggregation framework via between-subject attributes. Scandinavian Journal of Statistics. DOI: 10.1111/sjos.12695.
Liu, J., Zhang, X., Chen, T., Wu, T., Lin, T., Jiang, L., ... & Tu, X. M. (2021). A semiparametric model for between‐subject attributes: Applications to beta-diversity of microbiome data. Biometrics. 2021; 1– 13. DOI: 10.1111/biom.13487
Liu, J., Lin, T., Chen, T., Zhang, X., & Tu, X. M. (2022). On Semiparametric Efficiency of an Emerging Class of Regression Models for Between-subject Attributes. arXiv preprint arXiv:2205.08036. DOI: https://doi.org/10.48550/arXiv.2205.08036
Liu, J., Xu, K., Wu, T., Yao, L., Nguyen, T. T., Jeste, D., & Zhang, X. (2023). Deciphering the ‘gut–brain axis’ through microbiome diversity. General Psychiatry, 36(5). DOI: http://dx.doi.org/10.1136/gpsych-2023-101090
Nguyen, T. T., Zhang, X., Wu, T. C., Liu, J., Le, C., Tu, X. M., ... & Jeste, D. V. (2021). Association of loneliness and wisdom with gut microbial diversity and composition: an exploratory study. Frontiers in psychiatry, 395. DOI: 10.3389/fpsyt.2021.648475
Duan, Y., Llorente, C., Lang, S., Brandl, K., Chu, H., Jiang, L., White, R.C., Clarke, T.H., Nguyen, K., Torralba, M., Shao, Y., Liu, J., ... & Schnabl, B. (2019). Bacteriophage targeting of gut bacterium attenuates alcoholic liver disease. Nature, 575(7783), pp.505-511 DOI: 10.1038/s41586-019-1742-x
Lang, S., Duan, Y., Liu, J., Torralba, M. G., Kuelbs, C., Ventura Cots, M., ... & Schnabl, B. (2020). Intestinal fungal dysbiosis and systemic immune response to fungi in patients with alcoholic hepatitis. Hepatology, 71(2), 522-538. DOI: 10.1002/hep.30832
Featured Talks:
Dec. 1-2, 2023 – “Demystifying Causal Interpretations from High-dimensional Potential Outcomes: Applications to Microbiome Data.”- invited speaker, Interactive Causal Learning Conference, Boca Raton, Florida
Sept. 19, 2023 – “Feature Aggregation and Causal Discovery for High-dimensional Data: Application to Targeting the ‘Gut-Brain-Axis’ via Microbiome Diversity.” – YSPH Biostatistics Seminar, Department of Biostatistics, Yale University, New Heaven
June 11-14, 2023 – "Ensembled Distance-Based Generalized Estimation for Repeated Measures (Edger) Of Between-Subject Attributes in Longitudinal Microbiome Data." ICSA Applied Statistics Symposium, Ann Arbor
March 19-22, 2023 -- “Semiparametric Approaches for Inter- and Intra-Individual Variability through Generalized Distances of Wearable Data.” ENAR 2023, Nashville
Item Response Theory (IRT) for Scale Development
I am also deeply interested in applying Structural Equation Model (SEM) in Psychometrics. Measurements in behavior science, psychology, and psychiatry are usually multidimensional that are not easily captured by a single variable or item. To comprehensively account for different components, we usually adopt Item Response Theory (IRT). Scale trimming and validating are also essential for efficient and effective measurements, especially in the information age. We need to optimize a unified short scale that is coherent, comprehensive, valid, and reliable. This is by no means merely selecting scale items with stringent statistical criteria without accounting for inherent traits captured by each question. I worked closely with content experts to address such a challenge.
References:
Liu, J, Lee, E., Strong, D., Dilip V. Jeste and Tu, XM. (2023). Scale trimming and validating: An effective short-form of the UCLA loneliness scale. In preparation.
Jeste, D.V., Thomas, M.L., Liu, J., Daly, R.E., Tu, X.M., Treichler, E.B., Palmer, B.W. and Lee, E.E., (2021). Is spirituality a component of wisdom? Study of 1,786 adults using expanded San Diego Wisdom Scale (Jeste-Thomas Wisdom Index). Journal of Psychiatric Research, 132, pp.174-181.
Thomas, M., Palmer, B., Lee, E., Liu, J., Daly, R., Tu, X., & Jeste, D. (2021). Abbreviated San Diego Wisdom Scale (SD-WISE-7) and Jeste-Thomas Wisdom Index (JTWI). International Psychogeriatrics, 1-10.
Lee, E. E., Bangen, K. J., Avanzino, J. A., Hou, B., Ramsey, M., Eglit, G., Liu, J., ... & Jeste, D. V. (2020). Outcomes of randomized clinical trials of interventions to enhance social, emotional, and spiritual components of wisdom: a systematic review and meta-analysis. JAMA psychiatry, 77(9), 925-935.
DeGruttola, V., Nakazawa, M., Lin, T., Liu, J., Goyal, R., Little, S., ... & Mehta, S. (2023). Modeling homophily in dynamic networks with application to HIV molecular surveillance. BMC Infectious Diseases, 23(1), 656.
Featured Talks (Upcoming & Past):
March 10-13, 2024 – “Harmonizing Latent Structure and Network Modeling of Psychometrics Using a Novel Distance-Based Regression.” ENAR Spring Meeting, Baltimore
Dec.16-18, 2023 – “Evaluating Latent Structures in the Graphical Network Model: Visual Exploration and Hypothesis Testing.” CMStatistics 2023, Berlin, Germany (present virtually)
July 25-28, 2023 – “A Shortened Positive and Negative Symptom Scale (Panss) Harmonizing Classical Item Response Theory With the Perspectives From Network Approach.” The IMPS 2023 Annual Meeting, College Park, MD
Feb.19, 2022 – “Scale Trimming and Validating: An Effective Short-form of the UCLA Loneliness Scale.” – Psychiatry Grand Rounds, New York University (present virtually)
Robust Semiparametric Modeling of Multi-subject Attributes with Functional Response Models (FRM)
Functional Response Models (FRM) generalizes the classical GLM to model multi-subject functional responses, which admit a wide range of applications, including enhancing the Mann-Whitney-Wilcoxon rank-sum test in survey data to address the restrictive test of equal distributions, and extending the inverse probability weighting (IPW) into a rank-based statistic to handle confounders in causal inference. Their estimators from the U-statistics-based generalized estimating equations (UGEE) all enjoy nice consistency, asymptotically normality, and efficiency. Such semiparametric models are more robust than traditional likelihood-based parametric models by relaxing the nuisance parameter to be infinite-dimensional.
References:
Lin, T, Chen, T, Liu, J, Tu, XM. (2021). Extending the Mann-Whitney-Wilcoxon rank sum test to survey data for comparing mean ranks. Statistics in Medicine. 2021; 40: 1705– 1717. https://doi.org/10.1002/sim.8865.
J. Kowalski, S. Hao, T. Chen, Y. Liang, J. Liu, L. Ge, C. Feng, X. M. Tu. (2018). Modern variable selection for longitudinal semi-parametric models with missing data. Journal of Applied Statistics. 2018; 45:14, 2548-2562 https://doi.org/10.1080/02664763.2018.1426739.
Chen, Ruohui, Lin, Tuo, Liu, Liu, Liu, Jinyuan, Chen, Ruifeng, Liu, Chenyu, Zou, Jingjing, Natarajan, Loki, Tang, Wan, Tu, Xin. (2022). A double robust estimator for Mann Whitney Wilcoxon rank sum test when applied for causal inference in observational studies. Under Review.
Chen, Guanqing, Zhang, Jasen, Lin, Tuo, Chen, Tian, Wang, Hongyue, Zhang, Hui, Tu, Xin, Liu, Jinyuan, Kowalski, Jeanne. (2022). On the functional response models and its applications to the intraclass correlation. Under Review.
Lin, T., Liu, J, Niu, X., Wu, T.S., Zhang, J., Li, Y., Richardson, S., Stander, V., Chen, G., Vu, T., Tu, J., Chen, T., Wang, B., Feng, C., Zhang, X., Liu, L. and Tu, X.M. (2022). On outcome and sampling weights; An in-depth look at the dueling weights. Under Review.