This page contains toolboxes and other code written by myself or our group for our research and made available for others. Although we hope you find it helpful, please do not expect a significant amount of technical support or debugging.

• Kernel Density Estimation (KDE) Toolbox for Matlab
• A reasonably efficient implementation of spatial data structures for kernel or Parzen window density estimation and similar functions. Most work is done by a k-d tree data structure; versions of the "fast Gauss transform" and nearest neighbor searches are also included. Written as a Matlab class, with methods for many standard distributional operations (estimating entropy, evaluating, etc.)

• Anomaly and Event Detection in Count Data
• Adaptive Inference in Graphical Models
• Gaussian Process Regression with Time-shifts