saasy-ml-layer-pipeline

SaaSy ML Pipeline

The machine learning pipeline that invokes JSAT algorithms.

Install

From within saasy-ml-layer-pipeline:

$ mvn install

The jar application will be generated in the target folder.

Use

From within saasy-ml-layer-pipeline\target:

"$ java -jar saasyml-pipeline-0.1.0-SNAPSHOT.jar -thread [1 | true] -serialize [1 | true] -tests [1 2 3 4 | Classifier Cluster Outlier Regressor]

where:

• thread: to use threads in the pipeline.
• serialize: to serialize the models.
• tests: builds, trains, and performs inference with a model.

Available model options are Classifier, Cluster, Outlier, and Regressor:

• Classifier: builds and tests a LogisticRegressionDCD model.
• Cluster: builds and tests a FLAME model.
• Outlier: builds and tests a IsolationForest model.
• Regressor: builds and tests a MultipleLinearRegression model.

For more details about the algorithms, check the JSAT API documentation at http://www.edwardraff.com/jsat_docs/JSAT-0.0.8-javadoc/.

To change the algorithms used in the tests, modify SaaSyMLPipeLineTest.

The following is an example to build and test a regressor model:

$ java -jar saasyml-pipeline-0.1.0-SNAPSHOT.jar -tests 4

A logs folder, with .log files, will be created inside the folder from where the command above is run.