MLlib Machine Learning in Apache Spark

This paper present spark’s open-source distributed machine learning library.

Integrate ML with spark has benefits:

• Spark is designed with iterative computation, which facilicate the ML algorithm in nature.
• Low-level components improvement has performance gain in MLLIB
• Simplify the deployment

Features:

• fast, distributed implementations of ML algorithms

• algorithmic optimizations to reduce JVM garbage collection, communication costs,

• Pipeline API: It address the overhead of pipeline construction.

---

• ← Previous Post
• Next Post →