I always try to keep every statistical model concise and straightforward as possible, so that it can make a direct impact on current scientific research. Many projects I worked on involved collaborations with scientists. The summaries of the projects I worked on are provided below.
1. Permutation-based inference of spatially localized signals in longitudinal MRI data
Idea: How to use spatial information effectively to construct a powerful inference procedure, detecting signal clusters while controlling for family-wise error rate?
[Slides coming soon] [R codes]
Presentation(s):
2. Semiparametric modeling of time-varying activation and connectivity in task-based fMRI data (TVAAC)
Idea: What are the sources of temporal dynamics of the brain in task-based fMRI? Is it connectivity, activation, or both?
[Slides] [R codes]
Presentation(s):
3. Integrative factorization of bidimensionally linked matrices (BIDIFAC)
Idea: Factorizing the linked structure of omics data consisting of multiple platforms and multiple cohorts.
[Slides] [R codes]
Presentation(s):
4. Effective data augmentation method for right-censored health records data
Idea: How can we apply existing binary classification methods to censored data for risk prediction?
5. Adaptive SNP-set association testing in generalized linear mixed models with application to family studies
Idea: Developing a "good" statistical test for association between a phenotype and a set of genotypes, when samples are genetically correlated.
[Slides] [R codes]
Presentation(s):
Idea: How to use spatial information effectively to construct a powerful inference procedure, detecting signal clusters while controlling for family-wise error rate?
[Slides coming soon] [R codes]
Presentation(s):
- JSM 2020 (scheduled)
- ENAR 2020
- UMN Biostatistics Student Seminar
2. Semiparametric modeling of time-varying activation and connectivity in task-based fMRI data (TVAAC)
Idea: What are the sources of temporal dynamics of the brain in task-based fMRI? Is it connectivity, activation, or both?
[Slides] [R codes]
Presentation(s):
- JSM 2019
- SMI 2019
- ENAR 2019
- UMN Biostatistics Student Seminar
3. Integrative factorization of bidimensionally linked matrices (BIDIFAC)
Idea: Factorizing the linked structure of omics data consisting of multiple platforms and multiple cohorts.
[Slides] [R codes]
Presentation(s):
- JSM 2019 (by Eric Lock)
- ICSA 2019
- UMN Biostatistics Student Seminar
4. Effective data augmentation method for right-censored health records data
Idea: How can we apply existing binary classification methods to censored data for risk prediction?
5. Adaptive SNP-set association testing in generalized linear mixed models with application to family studies
Idea: Developing a "good" statistical test for association between a phenotype and a set of genotypes, when samples are genetically correlated.
[Slides] [R codes]
Presentation(s):
- ENAR 2018