Sentiment Analysis Classifier Microservices

These microservices expose a REST API for client to POST a sentiment analysis prediction request. The training notebooks can be found in https://github.com/anonoz/fyp-ipynb.

Script	Classifiers Available
skipgram.py	lstm - Long short-term memory gru - Gated recurrent unit cnn - Convolutional neural network
dbow.py	ffnn - Regular feedforward neural net svm - Support vector machine rf - Random forest

Docker

These microservices have been Dockerized for your convenience. Running the command below will fetch the Docker image from Docker Hub and start the microservice right away, making it available on port 6001/6002 on your machine:

# may need sudo to run
$ docker run -p 6001:6001 -it anonoz/fyp-skipgram:0.0.1

# if you want to run both at the same time, open a new terminal and run this
$ docker run -p 6002:6002 -it anonoz/fyp-dbow:0.0.1

If you want to let them run in background without taking up the terminal, replace -it with -d.

In case you want to build the Docker image on your own, run:

$ docker build -t [image_name] .

Then run the commands in the snippet above the above one, replacing anonoz/fyp-microservices with your own image name.

API

POST /predict

{
    "classifier": "lstm",
    "review": "This movie is pretty damn good! ..."
}

Parameters

Field	Description
classifier	Can be lstm, gru, cnn, ffnn, svm, rf
review	The movie review text body to be analyzed.