Example usage for weka.core.converters CSVLoader setNominalLabelSpecs

Introduction

In this page you can find the example usage for weka.core.converters CSVLoader setNominalLabelSpecs.

Prototype

public void setNominalLabelSpecs(Object[] specs) 

Source Link

Document

Set label specifications for nominal attributes.

Usage

From source file:sentinets.Prediction.java

License:Open Source License

public int setInstances(String inputFile) {
    //String[] nominalVals =  {"42:positive,neutral,negative"};
    String[] nominalVals = { CLASSINDEX + ":" + StringUtils.join(classNames, ",") };
    original = null;/*from  w w  w.ja va  2s  .c o  m*/
    try {
        System.out.println("[In Prediction] Loading instances. ");
        CSVLoader csvSource = new CSVLoader();
        csvSource.setSource(new File(inputFile));
        csvSource.setFieldSeparator("\t");
        csvSource.setNominalAttributes(CLASSINDEX + "");
        csvSource.setStringAttributes(stringAttr);
        csvSource.setNominalLabelSpecs(nominalVals);
        original = csvSource.getDataSet();
        unlabled = original;
        classProbIndex = original.numAttributes() - 1;
        //System.out.println(unlabled.toSummaryString());
        Remove r = new Remove();
        //r.setAttributeIndices("3-4,6,10-12,14");
        if (classifierType == MODELTYPE.SENTIMENT || classifierType == MODELTYPE.SENTIMENT_WORD
                || classifierType == MODELTYPE.CUSTOM) {
            if (showProbability) {
                /*
                Add afilter;
                afilter = new Add();
                afilter.setAttributeName("last");
                afilter.setAttributeName("prediction_prob");
                afilter.setInputFormat(original);
                original = Filter.useFilter(original, afilter);
                classProbIndex = original.numAttributes()-1;*/
            }
            if (classifierType == MODELTYPE.SENTIMENT) {
                r.setAttributeIndices("3,4,6,8,10-12,14,42,43,45-last");
                System.out.println("Filtering instances for SENTIMENT");
            } else if (classifierType == MODELTYPE.SENTIMENT_WORD || classifierType == MODELTYPE.CUSTOM) {
                r.setAttributeIndices(removeAttr);
                System.out.println("Filtering instances for SENTIMENT WORD");
            }
        }
        //r.setAttributeIndices("3-4,6,10-12,14,40-41,43-last");
        r.setInputFormat(unlabled);
        unlabled = Remove.useFilter(unlabled, r);
        //System.out.println(unlabled.toSummaryString());
        r = new Remove();
        //System.out.println(unlabled.toSummaryString());

    } catch (FileNotFoundException e) {
        e.printStackTrace();
        return 1;
    } catch (IOException e) {
        e.printStackTrace();
        return 2;
    } catch (Exception e) {
        e.printStackTrace();
        return 3;
    }
    int cIdx = unlabled.numAttributes() - 1;
    unlabled.setClassIndex(cIdx);
    System.out.println(
            "Class Attribute is: " + unlabled.classAttribute() + " at index: " + unlabled.classIndex());
    return 0;
}

---

From source file:sentinets.SentiNets.java

License:Open Source License

public void setInstances(String inputFile) {
    String[] nominalVals = { "15:e,p", "16:s,na_ns" };
    original = null;/*from   ww w  .jav a 2s .co m*/
    try {
        CSVLoader csvSource = new CSVLoader();
        csvSource.setSource(new File(inputFile));
        csvSource.setFieldSeparator("\t");
        csvSource.setNominalAttributes("15-16");
        csvSource.setStringAttributes("3,4,6,8,10-12,14");
        csvSource.setNominalLabelSpecs(nominalVals);
        original = csvSource.getDataSet();
        unlabled = original;
        //System.out.println(unlabled.toSummaryString());
        Remove r = new Remove();
        r.setAttributeIndices("3-4,6,10-12,14");
        r.setInputFormat(unlabled);
        unlabled = Remove.useFilter(unlabled, r);
        //System.out.println(unlabled.toSummaryString());
        r = new Remove();
        if (classifierType == E_P) {
            System.out.println("Filtering instances for E_P");
            r.setAttributeIndices("9");
            r.setInputFormat(unlabled);
            unlabled = Remove.useFilter(unlabled, r);
        } else if (classifierType == S_NS) {
            System.out.println("Filtering instances for S_NS");
            r.setAttributeIndices("8");
            r.setInputFormat(unlabled);
            unlabled = Remove.useFilter(unlabled, r);
        }
        //System.out.println(unlabled.toSummaryString());

    } catch (FileNotFoundException e) {
        e.printStackTrace();
    } catch (IOException e) {
        e.printStackTrace();
    } catch (Exception e) {
        e.printStackTrace();
    }
    int cIdx = unlabled.numAttributes() - 1;
    unlabled.setClassIndex(cIdx);
}

---

From source file:sentinets.TrainModel.java

License:Open Source License

public void setInstances(String inputFile) {
    String[] nominalVals = { "42:positive,negative,neutral" };
    ins = null;/*from w  w w  . jav a2s  .  co m*/
    try {
        CSVLoader csvSource = new CSVLoader();
        csvSource.setSource(new File(inputFile));
        csvSource.setFieldSeparator("\t");
        csvSource.setNominalAttributes("15-16");
        csvSource.setStringAttributes("3,4,6,8,10-12,14");
        csvSource.setNominalLabelSpecs(nominalVals);
        ins = csvSource.getDataSet();
        Remove r = new Remove();
        r.setAttributeIndices("3-4,6,8,10-12,14,40-41");
        r.setInputFormat(ins);
        ins = Remove.useFilter(ins, r);
        //System.out.println(unlabled.toSummaryString());
        r = new Remove();
        System.out.println(ins.toSummaryString());

    } catch (FileNotFoundException e) {
        e.printStackTrace();
    } catch (IOException e) {
        e.printStackTrace();
    } catch (Exception e) {
        e.printStackTrace();
    }
    int cIdx = ins.numAttributes() - 1;
    ins.setClassIndex(cIdx);
}

---