This template provides sentiment analysis algorithm RNN.
Follow installation guide for PredictionIO.
After installation start all PredictionIO vendors and check pio status:
pio-start-all
pio status
Copy this template to your local directory with:
pio template get ts335793/template-scala-spark-dl4j-word2vec-rnn <TemplateName>
Download en-parser-chunking.bin and place it in <TemplateDirectory>/src/main/resources/
.
You can import example training data from kaggle. It is collection of the Rotten Tomatoes movie reviews with sentiment labels.
In order to use this data, create new app:
pio app new <ApplicationName> # prints out ApplicationAccessKey
set appName in engine.json to ApplicationName and import data with:
python data/import_eventserver.py --access_key <ApplicationAccessKey> --file train.tsv
You can always remind your application id and key with:
pio app list
You might build template, train it and deploy by typing:
pio build
pio train
pio deploy
In order to send a query run in template directory:
python data/send_query_interactive.py
and type phrase you want sentiment to be predicted. The result will be predicted sentiment for the phrase.