Iris

Experiment to test loading and processing iris dataset in Rust and using the processed data in Python for Machine Learning tasks.

Method

Load and process the iris dataset with polars, a fast, Rust library for working with DataFrames.

DataFrame is split into features and labels, and labels are converted into categorical values. The result will be two dataframe (one for features & one for label).

A Dataset object is created with linfa and then split into training and validation set. The training set is trained on a linfa_trees::DecisionTree model, and predictions are made on the validation dataset to compute the confusion matrix, accuracy, precision, recall and other metrices.

classes    | setosa     | versicolor | virginica
setosa     | 3          | 0          | 0
versicolor | 0          | 4          | 0
virginica  | 0          | 1          | 7

Metrics:
[src/model.rs:75] &cm.accuracy() = 0.93333334
[src/model.rs:75] &cm.precision() = 0.93333334
[src/model.rs:75] &cm.recall() = 0.9583333
[src/model.rs:75] &cm.mcc() = 0.8994902

In practice, large dataframe can be loaded and processed with Rust's polars crate and then saved to disk or a data warehouse which can be cheaply used by python to train or perform other data science tasks.

Examples

There's an examples/ folder containing both Rust & Python notebooks.

Contribution

You are very welcome to modify and use them in your own projects.

Please keep a link to the original repository. If you have made a fork with substantial modifications that you feel may be useful, then please open a new issue on GitHub with a link and short description.

License (MIT)

This project is opened under the MIT which allows very broad use for both private and commercial purposes.

A few of the images used for demonstration purposes may be under copyright. These images are included under the "fair usage" laws.