lklee9/driftmapper

There are 2 projects in this repository: DriftAnalyse and DataMapper.

Drift Analyse

Manually calls the java code to map and analyse drift in a dataset

Usage

• Run the gradle task "fatJar"
• Call the .jar file with the following commands depending on your use case:

$ java -jar <path to .jar> arg1 arg2 arg3 arg4 arg5 ...

Comand line arguments:
analyse      subsetLength1,subsetLength2,... StartIndex1,EndIndex1,StartIndex2,EndIndex2 output_folder file1 file2 ...
stream       subsetLength1,subsetLength2,... windowSize1,windowSize2,...                 output_folder file1 file2 ...
stream_cont  subsetLength1,subsetLength2,... windowSize1,windowSize2,...                 output_folder file1 file2 ...
stream_chunk subsetLength1,subsetLength2,... groupAttIndex,groupSize1,groupsSize2,...    output_folder file1 file2 ...
moving_chunk subsetLength1,subsetLength2,... groupAttIndex,groupSize1,groupsSize2,...    output_folder file1 file2 ...

• See main.java file for examples

Experiment types

Analyse (analyse)

• Analyses the drift between 2 windows.
• Is able to output detailed analysis of the drift including the drift in the lielihood and posterior before averaging.

Stream (stream)

• Comapares the drift between multiple fixed window sizes back to back
• eg. Compares drift between 1-10 and 11-20 then the drift between 11-20 and 21-30

Stream Continuous (stream_cont)

• Compares the drift between 2 moving windows with fixed sizes
• eg. Compares drift between 1-10 and 11-20 then the drift between 2-11 and 12-21

Chunk (stream_chunk)

• Groups the data into different chunks based on a given attribute
• Groups chunks of back-to-back instances with the same value of the given attribute to group on
• Given a group/time-window size, n, compares the first n chunk to the second n chunks and so on
• eg. Compares Chunks 1-2 and 3-4 then compares chunks 3-4 and 5-6

Moving Chunk (moving_chunk)

• Same as experiment type Chunk
• Difference: given a group/time-window size, n, compares 2 moving chunk windows that moves by 1 chunk at a time
• eg. Compares Chunks 1-2 and 3-4 then compares chunks 2-3 and 4-5

Data Mapper

A web application that provide a GUI interface for Drift Analyse.

Usage

A copy of the fatJar of DriftAnalyse is already added in "datamapper/lib"

1. Open a terminal in the directory "datamapper"
2. run the command "sbt" or "sbt.bat" in windows
3. run the command "run" in the sbt prompt
4. go to localhost:9000 in a browser

driftmapper's Issues

Can't run imported Gradle/java/scala projects

I am new in java/gradle. I checked several simple tutorials about Gradle but I need to run a more complicated package. I imported the DriftMapper project zip file. I followed instruction points

1- Run the gradle task "fatJar"

using the command shell I changed the directory to driftanalyse (a subdirectory of the DriftMapper) and execute the command:

gradle fatJar

The output:
A problem occurred evaluating root project 'driftanalyse'.
Could not find method compile() for arguments [{group=nz.ac.waikato.cms.moa, name=moa, version=2016.04}] on object of type org.gradle.api.internal.artifacts.dsl.dependencies.DefaultDependencyHandler.

The imported project applies the old version of Gradle 3.3. It is no problem because the folder contained Gradle wrapper.

./gradlew fatJar

The output:
Error: Could not find or load main class org.gradle.wrapper.GradleWrapperMain

In the second section of the instruction entitled Data Mapper we can read that "A copy of the fatJar of DriftAnalyse is already added in "datamapper/lib" but I cannot execute this file (java -jar ). - maybe it doesn't make sense.

Next, I followed instruction points for Data Mapper:

1- Open a terminal in the directory "datamapper"

2- run the command "sbt" or "sbt.bat" in windows

The output:

Java HotSpot(TM) Client VM warning: ignoring option MaxPermSize=256m; support was removed in 8.0
Error: Could not find or load main class xsbt.boot.Boot