Intelligent Ruby: Getting Started with Machine Learning
summary
Claims that more data often beats fancier code, demonstrated by tossing orders of magnitude more text at off‑the‑shelf NLP algorithms and improving accuracy; word probabilities estimated from web hit counts; image reduction and recommendations built by running Singular Value Decomposition on matrices; linear separators (perceptron → SVM) work after simple feature transforms and are available via libsvm bindings; similarity derived from gzip compression; ensembles of many simple models beat single complex ones (Netflix prize >10%); runtime pain from a mislabeled 40 TB cluster drives practical engineering tradeoffs.
AI-written summary · about
chapters (11)
- 0:00Start
- 0:22Goals
- 5:54Data Input
- 8:21Data-Driven Learning
- 13:11Learning as Compression
- 14:50Classification
- 19:53Recommendations
- 21:55Image Compression Algorithm
- 23:49Clustering
- 32:27The Github Challenge
- 34:33Resources
related talks
46mPractical Machine Learning and Rails
51mMachine Learning with Ruby
27mFive machine learning techniques that you can use in your Ruby apps today By Benjamin Curtis Machine learning is everywhere these days
29mDiscover Machine Learning in Ruby
23mFrom Data to Recommendations: Building an Intelligent System with ROR
54mMachine Learning explained with Examples
37mMachine Learning, The Right Way
37mLearning Machine Learning