Sign in

Jonathan J. Park

@jonathanpark.bsky.social
888 followers 168 following 42 posts

Assistant Professor @UCDavis in Quant Psych Discrete-/Continuous-Time Dynamic Networks and Community Detection www.JonathanPark.dev

PostsRepliesMedia
Jonathan J. Park @jonathanpark.bsky.social · 25/09/2026
These results suggest that continuous-time subgrouping could help us identify shared patterns that are less dependent on when we happen to measure our subjects. Special thanks to co-authors Nathan Xin Mills, @harrietteriese.bsky.social, and @omidvebrahimi.bsky.social
010
Jonathan J. Park @jonathanpark.bsky.social · 25/09/2026
Our simulations showed that cts-gimme more consistently preserved the correct groupings across sampling rates than its discrete-time counterpart when subgroup differences were subtle.
110
Jonathan J. Park @jonathanpark.bsky.social · 25/09/2026
If we then cluster people using those parameters, individuals can appear similar even when their underlying dynamics are truly different. The groups we identify may therefore reflect our sampling schedule rather than the processes we're trying to understand.
110
Jonathan J. Park @jonathanpark.bsky.social · 25/09/2026
Discrete-time models are widely used to analyze psychological time-series but their dynamic parameters depend on the interval between measurements. Measure the same person more or less frequently, and the effects we estimate can change.
100
Jonathan J. Park @jonathanpark.bsky.social · 25/09/2026
New preprint! People with different underlying processes may appear similar simply because of when we measure them; which we refer to as “illusory homogeneity” We introduce cts-gimme to identify group-, subgroup-, and person-specific dynamics in continuous-time: osf.io/preprints/ps...
osf.io
OSF
1102
Jonathan J. Park @jonathanpark.bsky.social · 18/09/2026
These results could--in part--help to explain why centrality indices can sometimes be hit or miss in their utility across different applications and scenarios. Thanks to @omidvebrahimi.bsky.social, @mijke.bsky.social, and @dennyborsboom.bsky.social for help throughout the process!
010
Jonathan J. Park @jonathanpark.bsky.social · 18/09/2026
Specifically, we simulated graphs from scale-free, small world, and random degree distributions and used a cascading failure simulation to show that targeting the most and least central nodes (e.g., Degree, Strength, Closeness, and Betweenness) matters more for some distributions than others.
110
Jonathan J. Park @jonathanpark.bsky.social · 18/09/2026
New preprint! Psychometric networks are often summarized with centrality indices that are sometimes useful and other times not. We show that the "most central" node's influence depends heavily on the configuration of edges in the network; its degree distribution. osf.io/preprints/ps...
osf.io
OSF
2187
Jonathan J. Park @jonathanpark.bsky.social · 30/08/2026
Choosing a model ultimately depends on your analytic goals and the alignment of those goals with the actual data on hand or the feasibility of gathering that data. Happy to discuss and collab! Feel free to DM. Preprint out soon😉
010
Jonathan J. Park @jonathanpark.bsky.social · 30/08/2026
cts-gimme doesn't always win. The discrete-time S-GIMME outperforms cts-gimme when clustering on large effect sizes (i.e., a lag-1 equivalent of ~0.60) combined with slow sampling rates due to its direct modeling of the shocks. tl;dr: think more deeply about the time-scales of your processes.
100
Jonathan J. Park @jonathanpark.bsky.social · 30/08/2026
The figure shows that cts-gimme groups people into the same group across sampling rates of 0.50, 1.00, or 5.00 time-units even at effect sizes at a lag-1 equivalent of ~0.20. By contrast, S-GIMME would assign subjects to similar groups at some sampling rates and change assignment for others.
100
Jonathan J. Park @jonathanpark.bsky.social · 30/08/2026
In our cts-gimme work, we highlight an issue with discrete-time models. Namely, observed effects are dependent on sampling rate. Then, if you cluster on those effects, your clusters may be artifacts of sampling rate rather than truth. We dub this, "illusory homogeneity".
172
Jonathan J. Park @jonathanpark.bsky.social · 29/08/2026
Our preprint for cts-gimme will be out soon with specifics!
010
Jonathan J. Park @jonathanpark.bsky.social · 29/08/2026
You might notice that subgrouping is also included in the package name. We’ve been doing simulations to improve the clustering of GIMME by using a partition around medoids-based approach which improves subgroup accuracy and consistency across a wide range of sampling rates from 0.5 to 5.0
120
Jonathan J. Park @jonathanpark.bsky.social · 29/08/2026
The GIMME framework is a method for analyzing heterogeneous time-series while identifying common dynamic patterns within samples. Happy to announce that we’ve extended it to the continuous-time framework in our new R-package, ctgimme! cran.r-project.org/web/packages...
cran.r-project.org
ctgimme: Continuous-Time Subgrouping with GIMME
Estimates group-, subgroup-, and individual-level dynamic structures from multivariate intensive longitudinal data using continuous-time state-space models. The subgrouping procedure combines iterativ...
1166
Jonathan J. Park @jonathanpark.bsky.social · 23/12/2025
A holiday gift from @suryadyutibaral.bsky.social for modeling longitudinal data in a non-parametric framework using continuous-time functional data analysis.
030
Jonathan J. Park @jonathanpark.bsky.social · 03/10/2025
My work focuses on how we can identify and address undiagnosed heterogeneity in samples of heterogeneous time-series by drawing on techniques from graph theory and network analysis. I also have a line of work directly in network analysis using cascading failure models and fuzzy clustering methods.
010
Jonathan J. Park @jonathanpark.bsky.social · 03/10/2025
Wanted to announce that I will be recruiting graduate students this year in Quantitative Psychology here at UC Davis. If you are or know of any undergrads who are interested in dynamical systems and network analytic methods, please send them my way or get in touch!
173
Jonathan J. Park @jonathanpark.bsky.social · 29/09/2025
whether a failure of a specific vertex results in overwhelming failure throughout a graph or whether the pattern of failure is similar to any other randomly selected vertex.
000
Jonathan J. Park @jonathanpark.bsky.social · 29/09/2025
Adding to (1.), we conduct simulations where peripheral vertices are more influential to one another and ones where hubs are highly connected *and* influential. In the former, we can recreate the results from this paper and in the latter we find that the topology of the graph is a large player in
110
Jonathan J. Park @jonathanpark.bsky.social · 29/09/2025
2. We don't allow flipped vertices to come back into the system; so, failure in our simulations is permanent. This could be a time-scale difference in the mode of collapse. Some systems fail so quickly that other variables cannot react while in slower time-scales, this is more plausible
100
Jonathan J. Park @jonathanpark.bsky.social · 29/09/2025
Thanks for the shout out, @omidvebrahimi.bsky.social. This is an interesting paper! Our simulations are a bit different. So, the type of world our models describe are a bit different and I can describe below: 1. We assumed that the connections between vertices are weighted and heterogeneous.
120
Jonathan J. Park @jonathanpark.bsky.social · 18/12/2024
Thank you, Cam! 🥹 Helps to be a part of a stellar department too 😉
010
Jonathan J. Park @jonathanpark.bsky.social · 18/12/2024
“Among several undergraduates I've worked with over the years, he is definitely in the top 20 (N = 20)."
040
Jonathan J. Park @jonathanpark.bsky.social · 16/12/2024
Thanks, Siwei!
000
Jonathan J. Park @jonathanpark.bsky.social · 16/12/2024
I wouldn't have been able to complete this work without mentors and collaborators: - Sy-Miin Chow - Peter Molenaar - @fishingwithzack.bsky.social - Michael Hunter - @chadshenkphd.bsky.social - Michael Russell
030
Jonathan J. Park @jonathanpark.bsky.social · 16/12/2024
In the paper, we highlight the strengths of modeling in continuous-time and contrast it with modeling in discrete-time dynamic networks. We also highlight some key weaknesses in implementing an automated search of continuous-time dynamic networks RE: initial conditions and determining them sensibly
130
Jonathan J. Park @jonathanpark.bsky.social · 16/12/2024
My first paper as an Assistant Professor at UC Davis is out in the journal SEM! In the paper, we introduce a continuous-time extension to the GIMME model implemented in OpenMx. It uses iterative tests of modification indices to construct group- and person-specific dynamic networks!
tandfonline.com
Unsupervised Model Construction in Continuous-Time
Many of the advancements reconciling individual- and group-level results have occurred in the context of a discrete-time modeling framework. Discrete-time models are intuitive and offer relatively ...
55813
Reposted by Jonathan J. Park
Jessica D Ayers @jessicadayers.bsky.social · 26/11/2024
Where did all the premies go during COVID👶? In this preprint, we (me, @jonathanpark.bsky.social, & M. Cardwell) discuss how lockdown measures may have (unintentionally) changed the way that genetic conflict 🧬manifested & the frequency of some pregnancy 🤰complications osf.io/preprints/ps...
osf.io
OSF
152
Jonathan J. Park @jonathanpark.bsky.social · 15/11/2024
I think you’re in charge of adding people if you made the thread! I am also new and don’t know anything hahah
110
Jonathan J. Park @jonathanpark.bsky.social · 15/11/2024
Nvm RE: my earlier thread. It’s here! 🙌
110
Jonathan J. Park @jonathanpark.bsky.social · 15/11/2024
Thanks, Björn! Looks like I’m too late to jump on that starter pack haha!
000
Jonathan J. Park @jonathanpark.bsky.social · 15/11/2024
Have to ask while Starter Pack Mania is at its apex: Have we made a Quant Psych starter pack? If we have, I’d love to be added but if not perhaps we can get one going.
210
Jonathan J. Park @jonathanpark.bsky.social · 14/11/2024
Hello! I’m primarily doing work in dynamic network modeling and community detection. Could I be added to this? Thanks for putting this together!
010
Jonathan J. Park @jonathanpark.bsky.social · 25/10/2024
Lots of new followers today thanks to @omidvebrahimi.bsky.social! Hello everyone! I'm an Assistant Professor at @ucdavispsych.bsky.social. I'm a quantitative psychologist studying dynamic network models in discrete- and continuous-time and how we can find commonalities in person-specific dynamics.
0153
Reposted by Jonathan J. Park
Christopher Crawford @cmcrawford.bsky.social · 24/09/2024
We have a new preprint! I'm excited to share recent work related to modeling multiple-subject, multivariate time series. We extend the multi-VAR framework to allow for data-driven identification and penalized estimation of subgroup-specific dynamics. arxiv.org/abs/2409.03085
arxiv.org
Penalized Subgrouping of Heterogeneous Time Series
Interest in the study and analysis of dynamic processes in the social, behavioral, and health sciences has burgeoned in recent years due to the increased availability of intensive longitudinal...
182
Jonathan J. Park @jonathanpark.bsky.social · 09/09/2024
Thanks, Björn! Glad you liked it; hope the reviewers do too haha We really wanted to be clear during the empirical application that we didn't magically solve issues with starting values in continuous-time. These systems are just so much more sensitive than discrete-time ones.
010
Jonathan J. Park @jonathanpark.bsky.social · 06/09/2024
Special thanks to my dissertation committee members: @fishingwithzack.bsky.social, Mike Hunter, Chad Shenk, Mike Russell, and my advisors Sy-Miin Chow and Peter Molenaar for their help and guidance throughout my PhD. I could not have done this without all of you!
000
Jonathan J. Park @jonathanpark.bsky.social · 06/09/2024
We also tested ct-gimme on real-world data and comment on some issues that researchers fitting continuous-time models are still likely to face even with the automated behavior of ct-gimme.
100
Jonathan J. Park @jonathanpark.bsky.social · 06/09/2024
We evaluated the performance of what we're calling ct-gimme in simulations and found that it outperforms N = 1 fitting in continuous-time by leveraging information across the sample prior to individual model fitting as per traditional GIMME
100
Jonathan J. Park @jonathanpark.bsky.social · 06/09/2024
Hi folks! Sharing a pre-print that we just submitted from my dissertation work. We adapted and extended the GIMME framework for identifying group-level structure in person-specific dynamics to the continuous-time framework via modification indices osf.io/preprints/ps...
osf.io
OSF
241
Jonathan J. Park @jonathanpark.bsky.social · 06/09/2024
Thanks, Aaron!
010
Jonathan J. Park @jonathanpark.bsky.social · 22/05/2024
Successfully defended my dissertation. I’m a doctor now! Got knighted by my advisors, Sy-Miin Chow and Peter Molenaar with an engraved sword gifted to my by Alexis Santos. Lots of folks to thank but never enough space in these character limits.
030
Jonathan J. Park @jonathanpark.bsky.social · 14/02/2024
Lots of exciting news in the coming months. Stay tuned! First up, our new paper just came out with Sy-Miin Chow, @sachaepskamp.bsky.social, and Peter Molenaar for fitting N = 1 graphical VAR models and using cluster-based methods to identify subgroups of people who share similar dynamic patterns.
tandfonline.com
Subgrouping with Chain Graphical VAR Models
Recent years have seen the emergence of an “idio-thetic” class of methods to bridge the gap between nomothetic and idiographic inference. These methods describe nomothetic trends in idiographic pro...
050