Many donation sites popped up due to COVID-19 and Black Lives Matter, but it is difficult to tell which ones are scams and which ones are real.
What it does
The model takes the inputted description text and classifies it as fake or real based on the training data.
How we built it
We used Google Cloud AutoML Natural Language to take in a dataset of GoFundMes (and similar donation sites) and to create a model based on that.
Challenges we ran into
There is no existing dataset of fake and real GoFundMes and donation sites, so we had to create it ourselves. Also, we've never done ML before, so it was a huge learning experience. There was a lack of documentation.
Accomplishments that we're proud of
We are proud that we were able to create a dataset and use ML to train the model properly.
What we learned
We learned how to use ML since we've never worked with it before.
What's next for Scam Prevention AI
We want to train the model on a larger dataset so that its accuracy is even higher.
Thank you, Rachael and Vivek