Machine Learning Library (MLlib) Guide
MLlib is Spark’s scalable machine learning library consisting of common learning algorithms and utilities, including classification, regression, clustering, collaborative filtering, dimensionality reduction, as well as underlying optimization primitives, as outlined below:
- Data types
- Basic statistics
- summary statistics
- stratified sampling
- hypothesis testing
- random data generation
- Classification and regression
- linear models (SVMs, logistic regression, linear regression)
- naive Bayes
- decision trees
- ensembles of trees (Random Forests and Gradient-Boosted Trees)
- isotonic regression
- Collaborative filtering
- alternating least squares (ALS)
- Dimensionality reduction
- singular value decomposition (SVD)
- principal component analysis (PCA)
- Feature extraction and transformation
- Frequent pattern mining
- Optimization (developer)
- stochastic gradient descent
- limited-memory BFGS (L-BFGS)
MLlib is under active development.
The APIs marked
DeveloperApi may change in future releases,
and the migration guide below will explain all changes between releases.
spark.ml: high-level APIs for ML pipelines
Spark 1.2 introduced a new package called
spark.ml, which aims to provide a uniform set of
high-level APIs that help users create and tune practical machine learning pipelines.
It is currently an alpha component, and we would like to hear back from the community about
how it fits real-world use cases and how it could be improved.
Note that we will keep supporting and adding features to
spark.mllib along with the
Users should be comfortable using
spark.mllib features and expect more features coming.
Developers should contribute new algorithms to
spark.mllib and can optionally contribute
See the spark.ml programming guide for more information on this package.
MLlib uses the linear algebra package Breeze, which depends on netlib-java for optimised numerical processing. If natives are not available at runtime, you will see a warning message and a pure JVM implementation will be used instead.
To learn more about the benefits and background of system optimised natives, you may wish to watch Sam Halliday’s ScalaX talk on High Performance Linear Algebra in Scala).
Due to licensing issues with runtime proprietary binaries, we do not
netlib-java’s native proxies by default. To configure
netlib-java / Breeze to use system optimised binaries, include
com.github.fommil.netlib:all:1.1.2 (or build Spark with
-Pnetlib-lgpl) as a dependency of your project and read the
netlib-java documentation for
your platform’s additional installation instructions.
To use MLlib in Python, you will need NumPy version 1.4 or newer.
spark.ml package, please see the spark.ml Migration Guide.
From 1.2 to 1.3
spark.mllib package, there were several breaking changes. The first change (in
ALS) is the only one in a component not marked as Alpha or Experimental.
- (Breaking change) In
ALS, the extraneous method
solveLeastSquareshas been removed. The
analyzeBlockswas also removed.
- (Breaking change)
StandardScalerModelremains an Alpha component. In it, the
variancemethod has been replaced with the
stdmethod. To compute the column variance values returned by the original
variancemethod, simply square the standard deviation values returned by
- (Breaking change)
StreamingLinearRegressionWithSGDremains an Experimental component. In it, there were two changes:
- The constructor taking arguments was removed in favor of a builder patten using the default constructor plus parameter setter methods.
modelis no longer public.
- (Breaking change)
DecisionTreeremains an Experimental component. In it and its associated classes, there were several changes:
DecisionTree, the deprecated class method
trainhas been removed. (The object/static
checkpointDirparameter has been removed. Checkpointing is still supported, but the checkpoint directory must be set before calling tree and tree ensemble training.
PythonMLlibAPI(the interface between Scala/Java and Python for MLlib) was a public API but is now private, declared
private[python]. This was never meant for external use.
- In linear regression (including Lasso and ridge regression), the squared loss is now divided by 2. So in order to produce the same result as in 1.2, the regularization parameter needs to be divided by 2 and the step size needs to be multiplied by 2.
Previous Spark Versions
Earlier migration guides are archived on this page.