This repository contains the prototype implementation of the Hierarchical Distributed Stream Classifier (HDSM), which is a distributed data mining approach to classifying heterogeneous data streams.
If you have Docker installed, you can run the experiments contained
within this codebase by executing make jupyter, opening the returned
URL in a web browser, and executing the contents of the provided
Jupyter notebooks (This has only been tested on an Ubuntu 16.04 host
running Docker 17.05.0-ce).
Note that the results of each experiment are saved to disk to prevent the need to re-execute the experiments when re-viewing an experiment's results.
- Java (>= 1.8.0)
- Leiningen (>= 2.0)
- For ClojureCLR:
- Mono
- Nuget (apt-get install)
make test
Tests can also be run repeatedly from a Clojure REPL:
lein repl(use 'midje.repl)(autotest)
See Makefile commands
This project was originally constructed to run on both JVM and CLR platforms (in order to use different underlying classifiers), but has since become dependent on java packages.
ClojureCLR 1.9.0 was used because 1.7.0 triggers the wrong reader conditional (it fires on :clj, which is used for JVM Clojure - https://dev.clojure.org/jira/browse/CLJCLR-76), and because 1.8.0 does not run successfully on Mono.