InvestAID

Inspiration

Usage of multiple data resources like generative AI, social media trends and sentiments to predict financial trends was something that we never worked on. There is scope to innovation as we harness the power of emerging yet powerful generative AI.

What it does

We sell our idea based on the following 3 features:

Topic Classification X Sentiment Analysis X ChatGPT
Social Media Trends X Sentiment Analysis X ChatGPT
Hoverables: Data X ChatGPT

We combine multiple ML models to generate useful and more exhaustive insights to make predictions in the volatile market of investment.

How we built it

Used Flair custom training for 3 way sentiment classification on open source news and articles data
Used Semi-supervised SVC for 5 way topic classification by identifying topics with K-means and manually labelling 100 samples.
Proposed to use few shot meta-learning architecture to classify the ever growing verticals based on text, social media hashtags and posting user
Proposed to build a self-learning model which uses all the features combined and treats the stock value scraped from Google stocks as the target for the historical data trends and thus, predicts in the present.

Challenges we ran into

Integration, data gathering, data labeling and model training in 1 day

Accomplishments that we're proud of

We are proud of all the features as they are all our innovations and harness the powers of significant indicative data resources required in predicting financial trends.

What we learned

Integration of generative AI to make impact on financial businesses and to predict trends with more confidence

What's next for InvestAID

Currently valid for only the popular companies as all sorts of open-source data is available
Internal data from the companies would add more features
Features from company profiles and their history from GS database
Server Deployment using Docker
Subscribing to ChatGPT 4.0 for a better accuracy