0

我正在为我的研究开发一个原型，我有训练有素的模型，我想用我的测试集进行测试。

我正在使用 c# 来开发程序。这是代码

            Classifier cls = null;
            cls = (Classifier)weka.core.SerializationHelper.read("Multinomial.model");


            //sample arff file

            //@relation emotion_class
            //@attribute Text string
            //@attribute class_list {Happy,Fear,Anger,Sad,Neutral}
            //@data
            //" ang saya saya #PlanPhilippines team interviewing #Bopha survivors Mindanao. ",

            StringBuilder buffer = new StringBuilder(sample);
            BufferedReader reader = new BufferedReader(new java.io.StringReader(buffer.ToString()));
            weka.core.converters.ArffLoader.ArffReader arff = new weka.core.converters.ArffLoader.ArffReader(reader);
            Instances dataRaw = arff.getData();

            //converts tweets to string to word vector
            StringToWordVector filter = new StringToWordVector();
            filter.setInputFormat(dataRaw);
            Instances dataFiltered = Filter.useFilter(dataRaw, filter);
            dataFiltered.setClassIndex(dataFiltered.numAttributes() - 1);


            for (int i = 0; i < dataFiltered.numInstances();i++ )
            {

                Instance inst = dataFiltered.instance(i);

                double classified = cls.classifyInstance(inst);
               MessageBox.Show(dataFiltered.classAttribute().value((int)classified));
            }

我收到一个错误：无法规范化数组。总和是 NaN。在double classified = cls.classifyInstance(inst);

但是当我将字符串删除到词向量过滤器时..没有错误但它输出错误的结果它总是给出 1 或快乐。

谢谢！

4

0 回答 0