Intro / Hypotesis
As an avid news reader, I sometimes wander on the need to involuntarily absorb negativity from the news pomp. And how does receiving “bad” or “negative” news on daily bases affect the mental health. I am not an expert psychologist, but I have a slight tiny feeling that absorbing negativity on daily bases isn’t necessarily a good. thing. How if, we get an AI model to suggest what to read ?
Yes, i know it might sound silly, but let’s give it a try.
Huggingface models, transformers, tokenizers
To begin with, I thought let me start with a model that’s not too huge, yet quite decently accurate.
https://huggingface.co/nlptown/bert-base-multilingual-uncased-sentiment
Sentiment Analysis
Sentiment analysis is the process of tagging data according to the sentiment like positive/negative and neutral. It really has a lot of other uses. Let’s imagine you have a Youtube Channel and you perform sentiment analysis on the comments ? It will give you a good idea on where you are heading.
Github
The project can be found on Github: https://github.com/sebs-tech/sentiment-news-analysis
Docker
The Docker Image can be found on DockerHub: https://hub.docker.com/repository/docker/sebstech/google-news-sa/general
To run it, just type:
docker run -p 8080:8080 sebstech/google-news-sa
From here onwards
Couple of ideas on my todo list:
- Get also article images and combine them with titles in the sentiment analysis.
- Find a place where to run the docker all the time and maybe pull other interesting stats
- Implement the functionality of user feedback, to fine tune the NLP model
- Give the option of several NLP models
- UI! Better UI!