Applied math, scientific programming, data science.

Featured

# Positive-unlabeled learning

A subfield of semi-supervised machine learning, where the only labeled data points available are positive.

Skip to content
## About me

## Modeling temperature

## Dabbling in data science competitions

## Price bins

## My favorite math problem

## Justifying a random forest’s predictions

## Bitcoin projections

## AUC with SQL

## Machine learning snippets

## Backyard trig

## Update on predicting Boston health code violations

## Latitude and the timing of childbirth

## Traffic stops in Charlotte NC

Applied math, scientific programming, data science.

Exploring 17 years of Asheville temperature data

How I tackled a couple of recent machine learning contests.

Using a clustering technique to segment a set of products into natural groups.

No calculus required.

LIME-like explanations in a fraction of the time.

A ten-minute hypersimplistic analysis, just for fun.

Using basic SQL to find the area under the ROC curve.

Pieces of code to do a few neat things with machine learning models.

Using a bit of scientific computing to solve an inadvertently difficult math problem.

Reworked to use live data.

Trying to shed a little more light on a mysterious relationship.

A brief exploration of data from traffic stops, their participants, and their outcomes.