About me
Welcome! My name is Adrian Vrabie. I would like to share my interests for Data Science and Hidden Markov Models.
I consider myself fortunate to have enjoyed the works of Dmitrii Lozovanu on Markov Chains, Valentin Belousov on Quasigroups, Andrew Viterbi on Dynamic Programming and many others. All of them successfully humbled me, and I hope to help you find the beauty in them too.
Another part of me is an Economist by training. Economics is a social science that extensively uses the mathematics apparatus to provide plausible explanations. Ex: with few assumptions, Economics can explain why the Baum-Welch algorithm, despite its applicability and insight, is less popular than the Viterbi algorithm. Currently I am involved in the Energy sector: have been in the exploration phases of upstream petroleum and getting involved in lng terminals and logistics.
Before modelling, it is imperative for me to get some insider views and assess the kernel of each major project.
Although not as comprehensive as a CV, I've put some info that holistically reflect the amalgam of my interests in Applied Mathematics, Programming and Data Science.
- Dynamic Stochastic Systems and Systems of Differential Equations
- Discrete and Continuous State Markovian Processes
- Hidden Markov Models
- DSGE models
- Algorithms for Data Analysis
- Clustering and classification
- Regression models
- Logistic models
- Principal Component Analysis
- Matrix decomposition algorithms
- Programming:
- SQL
- Julia / Python
- Java
- C/C++
- HTML, CSS, JavaScrip, Bootstrap, JQuery, JavaScript
- Applications of Markovian processes in Economics, Finance and Genomics
Things I'm interested in
Portfolio
Leontief Matrix Project: for calculating the impact of an exogenous US$ on the economy.
A Gentle Introduction to Markov Chains, Hidden Markov Models and Baum Welch algorithm. (HTML) Latent Parameters Estimation
FreeCodeCamp project: A tribute to Leonard E. Baum and Lloyd R. Welch. (to be continued) link here
Fractal Geometry Project. Selected programmed exercises. Using a contractor we can easily construct its image using very little memory.
My old coding blog Inversing a Matrix using the Jordan Gauss method. Coded in Python, with easy to follow examples.
My Slides for the 2016 CECMI conference at the State University of Moldova. Methods for estimating the state transition matrix in Hidden Markov Models.
Publications
- Thesis A Vrabie "Applications of Homogeneous Markovian Processes in Economics and Latent Parameters Estimation for Discrete State Hidden Markov Models" State University of Moldova, 2016 Advisor: Phd. Habilitat Professor Dmitrii Lozovanu
- (forthcoming:) A Vrabie Methods in Estimating the State transition matrix in HMMs International Scientific Conference Mathematics & IT: Research and Education MITRE-2016
- A Vrabie "Economic Growth and Education in Republic of Moldova: from Illusion to Reality", Economica An XX nr 1 (79), March 2012, Academy of Economic Studies of Moldova pp 18-28
Games
I made a few since high school, here's one of them
