public static void AttributeSelection() {
// search sequences from mirbase based on a keyword
System.out.println("Attribute selection");
Mirbase m = new Mirbase();
ArrayList altrain =
m.getSequencesFromFiles(matures, hairpins, organisms, structures, keyword, "tmp");
// train model
File arff = new File(keyword + ".arff");
AdaptDataForWeka.createFileFromList(altrain, arff, bestFeatures);
System.out.println("Attribute selection for " + keyword);
WekaModule.attributeSelection(arff, keyword + ".AttribSel.txt");
}
/** Use miRbase fastas source */
public static void miRdupExecutionFastas() {
String filename = matures.substring(0, matures.lastIndexOf("."));
// Normal execution. We get sequences on mirbase using keyword ("all" if empty),
// then we train the model on it, and if the test dataset is present we submit it to the model
if (!classifier.isEmpty()) {
if (predictionsFile.isEmpty()) {
System.out.println("You must submit a file with data you want to validate");
} else {
System.out.println("Validation of " + predictionsFile + " with classifier " + classifier);
// test model
File arff = null;
if (arffFileForValidate.isEmpty()) {
arff = new File(predictionsFile + ".arff");
AdaptDataForWeka.createPredictionDataset(predictionsFile, false);
} else {
System.out.println("Using arff file " + arffFileForValidate + " to validate");
arff = new File(arffFileForValidate);
}
WekaModule.testModel(arff, predictionsFile, classifier, predictMirnaPosition);
}
} else if (trainFromOnlineMiRbase == true) {
// search sequences from mirbase based on a keyword
Mirbase m = new Mirbase();
ArrayList altrain = m.getSequencesFromMirbase(keyword);
// train model
File arff = new File(keyword + ".arff");
AdaptDataForWeka.createFileFromList(altrain, arff, false);
WekaModule.trainModel(arff, keyword);
// submit a file with predicted mirnas and their hairpins to the model
if (!predictionsFile.isEmpty()) {
// test model
arff = null;
if (arffFileForValidate.isEmpty()) {
arff = new File(predictionsFile + ".arff");
AdaptDataForWeka.createPredictionDataset(predictionsFile, false);
} else {
arff = new File(arffFileForValidate);
}
WekaModule.testModel(arff, predictionsFile, keyword + modelExtension, predictMirnaPosition);
}
// Case offline mirbase, and local sequences are submitted
} else {
// search sequences from mirbase based on a keyword
File arff = null;
if (arffFileForTrain.isEmpty()) {
Mirbase m = new Mirbase();
ArrayList altrain =
m.getSequencesFromFiles(matures, hairpins, organisms, structures, keyword, filename);
arff = new File(keyword + "" + filename + ".arff");
AdaptDataForWeka.createFileFromList(altrain, arff, false);
} else {
arff = new File(arffFileForTrain);
}
// train model
WekaModule.trainModel(arff, keyword + "" + filename);
// submit a file with predicted mirnas and their hairpins to the model
if (!predictionsFile.isEmpty()) {
// test model
arff = null;
if (arffFileForValidate.isEmpty()) {
arff = new File(predictionsFile + ".arff");
AdaptDataForWeka.createPredictionDataset(predictionsFile, false);
} else {
arff = new File(arffFileForValidate);
}
WekaModule.testModel(
arff, predictionsFile, keyword + filename + modelExtension, predictMirnaPosition);
}
}
}
