I’m excited to announce the v0.1.x of Canapes.jl, a pure-Julia package designed for statistical learning on sparse matrices. It brings matrix factorization, item-item similarity, low-rank completion, and sparse regression.
Recommender systems and collaborative filtering have lacked maintained, production-ready tooling with existing options often outdated or unmaintained. This library aims to close that gap by creating a high-performing library competing with (python) implicit, (R) rsparse and others well know libraries.
The package is still work in progress, so any feedback or contribution is very welcome. I hope it helps to mature the ecosystem for recommendation engines and we can take Julia to the next level!