README.md

Polars Resources

• Ruby Polars Repo/Docs
• Polars API reference
• Vega -- Graphing library to use with Polars

Social Communities

• X Ruby AI group
• Mastodon Ruby AI and Data group

Machine Learning/AI

• Awesome Machine Learning with Ruby repo
• Andrew Kane's ML Gems
• Landon Gray's introduction to Machine Learning with Ruby
• Justin Bowen's Discover Machine Learning in Ruby
• Machine learning Docker images for Ruby ml-stack
• Rumale (Ruby machine learning) is a machine learning library in Ruby. Rumale provides machine learning algorithms with interfaces similar to Scikit-Learn in Python.
• ONNX Runtime Ruby high performance scoring engine for ML models - for Ruby

Deep Learning

• Torch.rb and TensorFlow
• TorchVision for computer vision tasks
• TorchText for text and NLP tasks
• TorchAudio for audio tasks
• TorchRec for recommendations

Forecasting

• Prophet.rb is a forecasting library. It supports multiple seasonalities, holidays, growth caps, and many other features.
• MITIE does named-entity recognition. It finds people, organizations, and locations in text.

Other DataFrame Gems

• Daru
• Rover
• RedAmber

Jupyter Notebook

• IRuby
• Jupyter
• Google Colab
• Host Jupyter Notebooks online mybinder

N-dimensional Array Gems

• Numo::NArray
• NMatrix

Misc

• SciRuby
• Outlier Detection and Removal Though in Python it's a good high level introduction to the statistical concepts. Introduces you to IQR and Standard Deviation, the two main outlier removal types.

# Std Deviation - filtering 2 std deviations away
df = Polars::DataFrame.new({'example' => [10, 20, 30, 40, 50, 10000]})
mean = df['example'].mean
std = df['example'].std
outliers = (df['example'] < mean + (2 * std)) & (df['example'] > mean - (2 * std))
removed = df.filter(outliers)

#one liner
removed = df.filter((df['example'] < df['example'].mean + (2 * df['example'].std)) & (df['example'] > df['example'].mean - (2 * df['example'].std)))

#IQR
series = Polars::Series.new('example', [-2.08,  1.9, -5.70, 7.08,  0.73, -3.50, 2.57,  0.21, -9.26])
series.quantile(0.75, interpolation:"nearest")