tund/skvm

Sparkling Vector Machines

Sparkling Vector Machines

These codes implements Sparkling Vector Machines (SkVM) using MATLAB and Apache Spark with Python API. The codes are tested on MATLAB R2014b and Apache Spark version 1.4.1, Python 2.7. Please make sure that you have installed python-numpy and you can run Spark on local to run the demo.

Run a demo of SkVM for classification

MATLAB: run_skvm.m
Apache Spark: spark-submit --master local[*] skvm_spark.py

Run a demo of SkVM for label-drift classification

MATLAB: run_skvm_labeldrift.m
Apache Spark: spark-submit --master local[*] skvm_spark_labeldrift.py

Run a demo of SkVM for regression

MATLAB: run_skvm_regression_airline.m
Apache Spark: spark-submit --master local[*] skvm_spark_regression.py

Citation

Tu Dinh Nguyen, Vu Nguyen, Trung Le, Dinh Phung. "Distributed Data Augmented Support Vector Machine on Spark". Accepted for the 23rd International Conference on Pattern Recognition (ICPR), 2016.