hyperSMURF

Weka implementation of hyperSMURF using EasyEnsemble and SMOTE.

Please read the manual for more information, a detailed installation descriotion and a tutorial.

Requirements

hyperSMURF requires java 8 and higher. It can be used as a Weka plugin using version 3.9 or higher. To build the program it is recommended to use Maven.

Installation from Github sources

To use hyperSMURF in Weka three steps are needed.

Clone this repository.
Compile the java classes and create a Weka plugin using Maven
Load the plugin into your Weka Package Manager

Clone this repository

Use your terminal and go to a folder where you want to checkout hyperSMURF. Then run:

git clone https://github.com/charite/hyperSMURF.git

Compile the java classes and create the Weka plugin

Go to your repository and create a jar file of hyperSMURF using Maven.

cd hyperSMURF

mvn clean install package

Now you should have the hyperSMURF-0.3.jar in the folder target/. The package phase of Maven creates also the Weka file hyperSMURF-0.3-weka.zip. It is located in the target/ folder.

Load the plugin into your Weka Package Manager

Open Weka, go to the package manager, and load the file hyperSMURF-0.3-weka.zip into it. Look at the Weka wiki for more information about the Weka Package Manager.

Installation using Maven

If you want to include hyperSMURF into your java project you can use Maven to download the necessary files from Maven Central. You have to add this code under the dependencies section in your pom.xml:

<dependency>
    <groupId>de.charite.compbio</groupId>
    <artifactId>hyperSMURF</artifactId>
    <version>0.3</version>
</dependency>

Citation

Please cite our Scientific Reports article:

M. Schubach, M. Re, P. N. Robinson, and G. Valentini. (2017).
Imbalance-Aware Machine Learning for Predicting Rare and Common Disease-Associated Non-Coding Variants.
Scientific Reports, 7.

charite / hyperSMURF

readme