Issue with Using AventIQ for News Sentiment Analysis

Models
1

Hello, I am building an AI application to predict stock market movement.

I have already implemented Technical Analysis (TA) and Fundamental Analysis (FA) scoring, and both are working well.

Now I am trying to include news-based sentiment analysis. I am using the “GNews API” to collect news. Since my company is a startup, we cannot afford expensive paid APIs.

So I installed the “AventIQ” model for sentiment analysis of news. However, the model is only analysing the headline from GNews and not the full article content from the URL.

I want the model to take the full article body for analysis, not just the title.

Could you please guide me on how to make AventIQ analyse the full article text?

Also, if there are any other models on Hugging Face that can help with analysing the sentiment of full news articles, I would be grateful for your suggestions.

Thank you.

1 Like
2

There seem to be several models available for analyzing sentiment in finance and markets, so it might be worth looking into.

If you want to analyze the text, you can simply feed the text into the model, so there is no need for a specialized model. However, some models may be better suited for handling long texts than others…

Also, it seems that AvantIQ-AI has multiple models available, but which one are you referring to?
There may already be a model available from AvantIQ-AI or other providers that suits your needs.

3

Thank you for your reply and for sharing these resources.

I am currently using the AventIQ-AI/sentiment-analysis-for-stock-market-sentiment model. I fetch news using GNews, then try to analyse the sentiment. But the problem is that the model only processes the headline returned by GNews, not the full article content from the actual news URL.

I understand that I can feed the full text into the model manually, but I’m not sure of the best way to extract the full article body from the URL. If you know any tools or libraries to do that efficiently (for example, from a Python script), I would really appreciate your suggestions.

Also, I’ll definitely look into the other models you shared like FinBERT and the GitHub project for financial sentiment. Thank you again for pointing me in the right direction.

1 Like
4

If this is GNews, it seems that the instructions are in the manual.

from gnews import GNews

google_news = GNews()
json_resp = google_news.get_news('Pakistan')
article = google_news.get_full_article(json_resp[0]['url'])  # newspaper3k instance, you can access newspaper3k all attributes in article
print(article.text) # article body