bert-sentiment-analysis

BERT Sentiment Analysis Example Flow

This example flow uses a BERT Sentiment model to classify the sentiments of comments for a Youtube video and charts the result.

Install the Dependencies

In order to correctly load Tensorflow JavaScript npm package: @tensorflow/tfjs-node, make sure there is no other @tensorflow/tfjs-node package that could be searched by Node.js require() under node_modules directories in current directory and all its parent directories.

Then run the npm install to install all dependencies in current directory.

Launch Node-RED

Now you can use npm run start to launch the object detection flow and access the Node-RED editor in https://localhost:1880.

Walk Through the Details

In the main flow, comments are processed by the sentiment analysis subflow to be classified as positive or negative. The classification is then shown in a chart.

You may use inject to trigger the flow. In the inject and under payload, select JSON to modify the object that contains:

• video_id: the unique id can be found in a video uri prefixed with v=.
• max_comments: only pull the number of latest comments.

Example object: {"video_id":"9bZkp7q19f0", "max_comments":"100"}

The comments are pulled down by the Read Comments function node and then fed into the Sentiment Analysis subflow which basically sanitizes the comments, tokenizes them, and classifies sentiments.

The result is charted by the chart subflow and the graph can be accessed from the dashboard tab in the sidebar (or just localhost:1880/ui/).

The comments and classification score can be seen from the debug tab in the sidebar, as well.