syaroslavtsev / storm-pattern Goto Github PK

#Storm.pattern

This project is based on the cascading.pattern project. The pattern sub-project for http://Cascading.org/ which uses flows as containers for machine learning models, importing PMML model descriptions from R, SAS, Weka, RapidMiner, KNIME, SQL Server, etc.

All Credit to Chris and Paco for the excellent work!

Current support for PMML includes:

• Random Forest in PMML 4.0+ exported from R/Rattle
• Linear Regression in PMML 1.1+
• Hierarchical Clustering and K-Means Clustering in PMML 2.0+
• Logistic Regression in PMML 4.0.1+

First include the clojars repo in your POM (or project.clj or sbt or wherever):

<repositories>
	<repository>
		<id>clojars.org</id>
		<url>http://clojars.org/repo</url>
	</repository>
</repositories>

And then add the dependency:

<dependency>
	<groupId>com.github.quintona</groupId>
	<artifactId>storm-pattern</artifactId>
	<version>0.0.2-SNAPSHOT</version>
</dependency>

I have created a very simple trident topology to illustrate the usage, it is available from here.. At a high level, this is all that is required:

topology.newStream("valueStream", spout)
			.each(new Fields(fields), new ClassifierFunction(pmml_file),
					new Fields("prediction"))
			.each(new Fields("prediction"), new PrintlnFunction(),
					new Fields());

You simply need to create the Classifier function and pass in the model.

To build and then run its unit tests:

mvn clean install

The following scripts generate a baseline (model+data) for the Random Forest algorithm. This baseline includes a reference data set -- 1000 independent variables, 500 rows of simulated ecommerce orders -- plus a predictive model in PMML:

./src/py/gen_orders.py 500 1000 > orders.tsv
R --vanilla < ./src/r/rf_pmml.R > model.log

This will generate huge.rf.xml as the PMML export for a Random Forest classifier plus huge.tsv as a baseline data set for regression testing.

Check the src/r/rattle_pmml.R script for examples of predictive models which are created in R, then exported using Rattle. These examples use the popular Iris data set.

• random forest (rf)
• linear regression (lm)
• hierarchical clustering (hclust)
• k-means clustering (kmeans)
• logistic regression (glm)
• multinomial model (multinom)
• single hidden-layer neural network (nnet)
• support vector machine (ksvm)
• recursive partition classification tree (rpart)
• association rules

To execute the R script:

R --vanilla < src/r/rattle_pmml.R

It is possible to extend PMML support for other kinds of modeling in R and other analytics platforms. Contact the developers to discuss on the cascading-user email forum.

• Data Mining Group XML standards and supported vendors
• PMML In Action book
• PMML validator