March 06, 2017
Hyperparameter tuning is an important step for maximizing the performance of a model. Several Python packages have been developed specifically for this purpose. Scikit-learn provides a few options, GridSearchCV and RandomizedSearchCV being two of the more popular options. Outside of scikit-learn, the Optunity, Spearmint and hyperopt packages are all designed for optimization. In this post, I will focus on the hyperopt package which provides algorithms that are able to outperform randomized search and can find results comparable to a grid search while fitting substantially less models.
December 16, 2016
While I was a teaching assistant for an undergraduate numerical analysis course, I developed this two-page Matlab cheatsheet. It covers basic programming concepts in Matlab and provides an example on generating a plot. This cheatsheet is a valuable resource when first learning Matlab.
November 30, 2016
Bohemian eigenvalues are the distribution of the eigenvalues of a random matrix. More specifically, this article explores matrices of low dimension (typically no larger than 10 x 10) whose entries are integers of bounded height. The name "Bohemian" is intended as a mnemonic and is derived from "Bounded Height Integer Matrix Eigenvalues" (BHIME). Here I present an overview of some preliminary results arising from this project.
November 23, 2016
Matlab provides a powerful plotting interface. Unfortunately, the default settings do not produce the most visually appealing figures. Unlike other popular plotting libraries (ggplot2
, etc.), Matlab has no simple method for changing the overall appearance of a plot. One option is to use a platform such as Plotly within Matlab (more details here
). In this article I discuss how you can improve the appearance of the native plotting architecture within Matlab to give your figures a more professional, refined look.
November 23, 2016
Have you ever applied custom styling to a figure in Matlab only to discover this styling does not transfer to subsequent figures? Unfortunately, plot styling only applies to the figure it was initially applied to. Luckily, an under-documented feature in Matlab allows you to set global properties for your figures.