more unfinished work… SCSES
This method could be used in the case where we have two design matrices, and we’d like to see association of groups of both design matrices covariates (groups) with an outcome of interested (whatever, likelihood of purchasing an elephant). For some reason group lasso wasn’t working, either issues with software or some issue with the…
Gitting Started
If you’re a developer and you’re not using git for version control, you’re wrong. Moreover, git can, and should be used when developing statistical or machine learning models, for example if developing a “customized” Bayesian model in Stan or developing your own neural network architecture in TF/PyTorch, the latter I’m not as familiar with. Incrementalism.…
Staring at Trace Plots
I remember sitting in her office, with Eunjee Lee, PhD. We’re looking at a trace plot, this is a Metropolis-Hastings embedded in a Gibbs sampler, and there were multiple. We call it a “Metropolis-within-Gibbs.” So we’re looking at a trace plot. And she says zoom in, zoom in. So I zoom in. And she pauses,…
more litter from my trail of tears
Feel free to correct me or reply if there’s any obvious mistake I’m making. I might be telling on myself here. So during my masters, I was assigned an adviser who then assigned me a project. This was a research based masters with graduate level mathematics course requirements and a minimum 70 page thesis, in…
It’s been a while…
It’s been a while since I’ve posted here, mostly since I’ve been working for The SAS Institute Inc, as a software developer in the scientific computing department. We worked closely with the econometrics and time series engineering team. I want to say thank you, I couldn’t have asked for a better position after finishing my…
Imitation: Radar Waveform Design in a Spectrally Crowded Environment via Nonconvex Quadratic Optimization
On a whim, I chose to study Information and Communication Engineering, closest related keyword, Signal Processing, and Sister major, Electrical Engineering. I was paired with Cui GuoLong and his post-doc Xiang Xianyu. In order to learn more about radar filter design, I thought it would be a great idea to imitate one of his papers.…
Code for the Logo
Prior to starting at SAS, I was still consulting and working for Michigan. I wanted to make a logo for Likely LLC, to make it appear more credible. Models are meaningless without good advertising and visualizations, to non-statisticians. I was attempting to make an L with mathematics and machine learning/statistics. So I simulated a Paul…
Gaussian Process Model Dump, Aalto University Internship, Summer 2018 Part 2
This is a collection of models that I frequent if I’m doing any modeling with Gaussian processes. Contains logistic regression models with Gaussian process priors, spatial models, survival models, regression models with separate length scales, etc. There’s been some changes to the language, so you need to replace the name of the covariance functions to…
Models, Algorithms and Software for Statistical Inference
I’m often asked by recruiters, clients, academics or colleagues which models I’ve worked with. This post is an effort to put this all in one place. All of these models, algorithms, and software I’ve worked with in some capacity. This includes implementation, development, application, required understanding, or any combination thereof. This list in non-exhaustive. I’ll…
So What Actually Happened in Finland? Aalto University Internship Summer 2018, Part 1
After working in the psychiatry department at Michigan, I worked at Aalto University in Espoo, Finland next summer. There, I worked on the Stan math library implementing Gaussian process covariance functions and matrix utilities to make Gaussian process models more feasible in Stan. The ultimate goal was to implement what’s known as “the birthday problem”…
Clustering Using SVD on Omics Data
For Michigan, I’ve since applied the SVD based sorting algorithm, which can be used for clustering, to real data. The goal is to in some way model the relationship of proteins and metabolites, come up with modules, or groups of proteins and metabolites that were related, and then use these later in a regression model…
Multilevel Gaussian Processes to Model Hit Speed in Baseball
This is a practice test I was given prior to landing a contract with the Tampa Bay Rays from July 2021. Here, I was asked to predict hit speed given angle off the bat. There’s a publicly available github repository with the same data here: https://github.com/danhogan/batted-ball. Below is copied from the practice test. This is…
Additive Gaussian Process Time Series Regression in Stan
I’ve copied this over from discourse.mc-stan.org, but this was my post, so I’m comfortable doing so. While working in the psychiatry department at Michigan, I played around with EEG data. Next, I became curious about how to extract out different periodic components of a time series. I ended up finding a blog post on Andrew…
My Work at the University of Michigan Psychiatry Department in 2017, Part 1
After I finished my undergraduate degree at the University of Michigan in 2017, in the Mathematical Sciences, and Statistics, I worked in the University of Michigan Psychiatry Department. This was with Daniel Kessler, Dr. Eunjee Lee, Dr. Chandra Sripada, and Mike Angstadt. This was in 2017, so thank you for understanding if vocab isn’t up…
Clustering Using SVD
Dr. Murthy and I have been working on a way to interpret omics data. We’d like to see if there’s any natural grouping structure to a large dataset. We’ve computed something resembling a cross covariance matrix. After Dr. Murthy experimented with sparse CCA for a while, we tried some other clustering methods or regularization methods…
likely llc 