Presenting my research on fraud detection with Machine Learning

What did you use as sample data? Have you had the chance to use it on any live case? :)

@Python we have a live app on Playstore that uses the model and collects the data for us. I have a rant about the user metrics of the app reaching 10k users per month. Since we provide short term mobile loans, we use the data collected by the app

@IllSlapU Yes. Isn't that the whole point of machine learning? You want to train your model on what?? 🤷

@Yamemori no.no. our data usage is purely for mobile banking and ethical. Since i am the one who built the data pipeline i can assure you that no one else has access to that data and we will NEVER share

@CodesNotHot Thanks! Sounds really cool to be able to develop it with that kind of data :)

@IllSlapU
Collecting data is not a bad thing at all. The problems rise when you collect data in a way that you can link it to certain people or so that you can identify the people by the data you collected.
As long as data is collected completely anonymously, there is no real issue as you cannot link the information to the people the data was collected from.

@topic:
Really interesting. You have some link for a read up?

@DelError You should really check Computerphile's video on "The Anonymisation Problem"...

Paper? Is it still a thing?

Is it open?

@stacked Yeah. I am working on a paper for a local science paper and arXiv

@SuryaK Unfortunately the code is not because it relies heavily on private data and business models for the company. The algorithms will however be published very soon once i formalize everything