LaSP stands for Learning and Signal Processing. It provides an educational environment to learn how to use some of the awesome open source machine learning and signal processing packages out there. The tutorials are based on IPython. Installation dependencies can be managed through a virtual environment to simplify your experience.
The plan is to cover the following topics by Fall 2014:
- Introduction to Machine Learning
- Probability and Maximum Likelihood
- Gradient Descent
- Linear Models
- Neural Networks
- Principle Components Analysis
- Independent Components Analysis
- Clustering: k-Means, Gaussian Mixture Models, and Spectral Graph Methods
- Time Series - Covariance and Correlation Functions
- Time Series- Fourier Transform and Time-Frequency Representations
- Time Series - Coherence
- Singular Value Decomosition
- Sparse Dictionaries
We are in the process of projectizing old code, trying to follow the best practices detailed here. Some of the steps we need to take are:
- Get virtual-environment setup documented and working automatically
- Get setup.py filled in
- Document ALL functions!
- Get unit tests running from setup.py
- Implement IPython notebook for intro to ML
Coming soon!
Coming soon!