FinScore ML – pretrained models for text classification
Piotr Stepinski
Like this project
Posted Jan 22, 2025
We solved a multilingual transaction classification problem for money lending risk scores, using a pretrained Bert model to classify financial transactions.
To calculate the money lending risk score for users, we solved the financial clients' transaction classification problem for a multilingual case with no available training data for new languages. To achieve this, I incorporated a pretrained Hugging Face model.