Re: No prediction obtained

Did this use to work?

The output of

e.printStackTrace();

in your program would be helpful.

Cheers,
Eibe

On Fri, 24 Dec 2021 at 15:17, Bob Matthews <[email protected]> wrote:
>
> Hi
>
> I have just noticed that I am not getting a weka prediction !
> i.e. I get "Failed to make prediction" message ! (and I am unsure why)
>
> In my main program I have the following code:-
>
>      if ((DCC_conf_small_down) || (DCC_conf_small_up)) {
>
>            // run Weka, create new Instance and make prediction for
BBTheta
>            WekaApp wekaApp = new WekaApp();
>            // this should only be called once (or whenever the model has
changed and needs reloading)
>            try {
>                  wekaApp.init();
>            }
>            catch (Exception e) {
>              System.err.println("ERROR init: " + e);
>              e.printStackTrace();
>            }
>            // make a prediction
>            try {
>              int pred = wekaApp.currentInstance(OSV);
>              if (pred != -1)
>                myConsole.getOut().println("Predicted label: " +
wekaApp.getLabel(pred));
>              else
>                System.out.println("Failed to make prediction!");
>            }
>            catch (Exception e) {
>              System.err.println("ERROR predict: " + e);
>              e.printStackTrace();
>            }
>              myConsole.getOut().println("Weka finished");
>
>
>
> OSV is a double variable and is calculated prior to the above code
>
> At the end of my main code I have the following class:-
>
>   public class WekaApp {
>        /** the model to use. */
>        protected Classifier cls;
>      /** the training data header. */
>      protected Instances header;
>      /**
>       * Load model and header.
>       *
>       * @throws Exception if deserialization fails
>       */
>       public void init() throws Exception {
>           // [A] load model and header of data
>            // deserialize model
>            Object[] objs = SerializationHelper.readAll("C:/TSFDC Trading
2022/SimpleLogistic.model");
>            cls = (Classifier) objs[0];
>            header = (Instances) objs[1];
>       }
>
>       /**
>       * Returns the class label associated with the index.
>       *
>       * @param index the class label index
>       * @return the label
>       * @throws IllegalStateException if no header present
>       */
>       public String getLabel(int index) throws IllegalStateException {
>           if (header == null)
>               throw new IllegalStateException("No model/header loaded?");
>               return header.classAttribute().value(index);
>       }
>
>       /**
>       * Performs a prediction.
>       *
>       * @param v the value of the input variable to use for making a
prediction
>       * @return the predicted value (class index), returns -1 if it fails
to make prediction
>       * @throws Exception if prediction fails
>       */
>       public int currentInstance(double v) throws Exception {
>           try {
>               // [B] create instance
>               // create the current instance
>               double[] values = new double[header.numAttributes()];
>               for (int i = 0; i < values.length; i++)
>                   values[i] = Utils.missingValue();
>                   // - numeric
>                   values[0] = v;
>                   DenseInstance currentInstance = new DenseInstance(1.0,
values);
>                   currentInstance.setDataset(header);
>
>               // [C] classify nominal class
>               // classifyInstance() just returns the index of the
predicted label (the one with the highest probability) as a double
>               double prediction = cls.classifyInstance(currentInstance);
>               System.out.println("prediction: " + prediction);
>
>               return (int) prediction;
>           }
>           catch (Exception e) {
>               System.err.println("Weka exception occurred: " + e);
>               e.printStackTrace();
>               return -1;
>           }
>       }
>    } // end of class WekaApp
>
> Any suggestions as to why - no prediction ?
>
> Bob M
>
> _______________________________________________
> Wekalist mailing list -- [email protected]
> Send posts to [email protected]
> To unsubscribe send an email to [email protected]
> To subscribe, unsubscribe, etc., visit
https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
> List etiquette:
http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

_______________________________________________
Wekalist mailing list -- [email protected]
Send posts to [email protected]
To unsubscribe send an email to [email protected]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html