I have been playing with the Sentiment Classification using Word Vectors exmaple and got is to use my data and produce acceptable results. But my Question (not beeing an expert) is:
No I was to feed in different data using the model I trained. Where and how do i feed it in? Do I need to save the vector model? Any guideance would be appreciated!
If you have trained your model, you can feed in new data using the same input port of the Predictor node you use to feed in your test data. Of course, your new data needs to be preprocessed in the same way as your original data.
with the new release 3.3 next week there will be a node release that adjusts the feature space of document vectors. If you create vectors and train a pred. model with these vectors and you create a second set of vectors (of course using the same preprocessing pipeline) you can adjust the feature space of the second vector set according to the feature space of the first vector set using this new node.
I tried to modify the workflow 08_Sentiment_Classification_Using_Word_Vectors in order to apply the word vector to new Test data. But I don't understand why the algorithm classify everything wrong.
Please find attached the workflow. What am I missing? What is wrong?