How to Machine Learn

https://www.blogger.com/blogger.g?blogID=2678369358515250290#editor/src=dashboard

How to Machine Learn

step 0: Decide on a project that will force you to put everything together, this was mine

step 1: Install Anaconda (2.7) and Learn Pandas

• pandas cookbook
• things I wish I knew about pandas early on
• 10 min to pandas

step 2: Get Comfortable with ML Theory & PyData

https://www.coursera.org/course/ml

• Watch videos from coursera on logistic regression, support vector machines, clustering, and dimensionality reduction. The other videos are great too.

complement videos with these notebooks:

• supervised and unsupervised learning
• dimensionality reduction
• Practical Data Science
• General Assembly Content - https://github.com/justmarkham/DAT3
• More cool notebooks
• make nbviewer.ipython.org your friend (anytime you see a notebook on Github that looks cool paste the link into nbviewer)

step 3: Go Deeper

1. Linear Algebra
• focus on the basics and apply everything you learn within the context of numpy
• https://github.com/rougier/numpy-100
• http://www.engr.ucsb.edu/~shell/che210d/numpy.pdf
• http://setosa.io/ev/eigenvectors-and-eigenvalues/
• videos from Andrew Ng's course are great, but deeper knowledge will be helpful
• http://betterexplained.com/articles/linear-algebra-guide/
• https://www.coursera.org/course/matrix
• http://machinelearningmastery.com/linear-algebra-machine-learning
2. Deep Learning
• https://www.coursera.org/course/neuralnets
• http://www.cs.toronto.edu/~rgrosse/csc321/
• http://vision.stanford.edu/teaching/cs231n/
• dive into theano and apply it to a natural language processing problem using recurrent neural networks
• https://github.com/goodfeli/theano_exercises
• http://nbviewer.ipython.org/github.com/craffel/theano-tutorial/blob/master/Theano%20Tutorial.ipynb
• theano workshop on computer vision with convolutional neural nets
• https://www.youtube.com/watch?v=S75EdAcXHKk
• theano workshop by Frédéric Bastien
• https://github.com/gwulfs/nextml2015/tree/readme
• http://databoys.github.io/Feedforward/
• http://underflow.fr/ai/lets-play-with-theano-547
3. Play with Data
• Kaggle
• Open Data
4. Blogs, Research Papers, Podcasts, Meetups
• mlwave.com
• fastml.com
• reddit.com/r/machinelearning
• datatau.com
• startup.ml/resources
• thetalkingmachines.com
• partiallyderivative.com
• meetup.com/bostonml & meetup.com/sfmachinelearning
• Next.ML
• sebastianraschka.com
• honnibal.wordpress.com
• peekaboo-vision.blogspot.com
5. Web
• deploy your models with Python Microframework: Flask
6. More Topics
• NLP (RNN, LDA, & Word2Vec)
• https://github.com/IndicoDataSolutions/Passage
• http://radimrehurek.com/gensim/models/ldamodel.html
• http://radimrehurek.com/2014/02/word2vec-tutorial/
• Online Learning (Vowpal Wabbit)
• http://fastml.com/vowpal-wabbit-eats-big-data-from-the-criteo-competition-for-breakfast/
• https://github.com/JohnLangford/vowpal_wabbit/wiki/Tutorial
• http://zinkov.com/posts/2013-08-13-vowpal-tutorial/
• Graphical Models (HMMs)
• http://www.autonlab.org/tutorials/hmm.html
• Structured Prediction (PyStruct)
• https://pystruct.github.io/
• Ensemble Methods (Gradient Boosting & RFs)
• http://www.quora.com/How-do-random-forests-work-in-laymans-terms
• Hyper Parameter Optimization
• https://github.com/hyperopt/hyperopt
• Distributed Systems (Hadoop Ecosystem)
• http://web.stanford.edu/class/cs246/
• PySpark
• https://spark.apache.org/examples.html
• https://www.edx.org/course/scalable-machine-learning-uc-berkeleyx-cs190-1x#.VOC70VPF_lQ
• GPU Acceleration (Theano)
• http://deeplearning.net/software/theano/tutorial/using_gpu.html
• Computer Vision (CNN)
• http://deeplearning.net/tutorial/lenet.html
• Time Series
• http://scikit-learn.org/stable/modules/gaussian_process.html
• Visualization (tsne)
• https://github.com/karpathy/tsnejs
• http://bokeh.pydata.org/en/latest/
• http://stanford.edu/~mwaskom/software/seaborn/

• Reach out with Questions & Suggestions Here and remember: