vangj / ds

No tears data science tutorials

Intro

This repository is just some "no tears" tutorials on data science topics. You will need Jupyter to get the examples to run. The easiest way to get Jupyter is to install Anaconda as described below.

Conda

To create a Conda environment, you need to install Anaconda. Note, for Windows, do not type in source activate ds but instead activate ds. Assuming you are in the cloned ds repository directory, you may type in the following.

conda create -n ds python=3.6
source activate ds
pip install -r requirements.txt
python -m ipykernel install --user --name ds --display-name "ds"

Note that LibROSA requires a backend like ffmpeg. You will need to install that separately.

To remove the environment.

conda env remove -n ds --yes

Notebooks

• Solve logistic regression problems with gradient descent through Autograd, no tears
• Solve poisson regression problems with gradient descent through Autograd, no tears
• Solve regression problems with gradient descent through Autograd, no tears
• Dirichlet-Multinomial distribution, no tears
• Generating normally distributed values, no tears
• Random Bayesian Network generation, no tears
• Gradient descent, no tears
• Kullback-Leibler divergence, no tears
• Latent semantic analysis, no tears
• Neural network, no tears
• Psuedo R^2 for logistic regression, no tears
• Regression with dummy variables, no tears
• Using a Recurrent Neural Network (RNN) to classify sine wave signals, no tears
• Neural Networks, gradient descent, and regression, no tears
• Restricted Boltzmann Machine
• Audio processing and classification

Misc

To convert a ipynb to html, type in the following jupyter nbconvert --to html --template basic [file.ipynb]. To make the generated html file compatible with WordPress, type in sed -e 's/\$\([^$]*\)\$/$latex \1$/g' [file.html] > [output.html].

Copyright Stuff

Copyright 2018 Jee Vang

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

ds's Issues

When to stop for SGD algorithm?

Hi, I learn't a lot going through your GD implementation steps described in gradient-descent.ipynb. Thank you for sharing the information.

For batch gradient descent implementation, you have used 10,000 iterations. I was wondering how did you decide on the # of iterations to be 2000 for SGD?