How did you successfully apply machine learning in a company? What impact do machine learning have for businesses?
Many people question the effectiveness of machine learning in helping business improve.
Here we share the impact that deployed machine learning systems made on business related metrics. You are free to send anonymous results.
This is a work in progress, feel free to suggest changes and contribute!
Simple description of the task
Metric: average order size
Benchmark: best selling products of the past four weeks
Difference: +20%
We wanted to rank the news articles according to the previous behavior of our visitors
Metric: average time spent on website
Benchmark: most shared articles on social media
Difference: +12.5%
They charged five cents for each transaction, but since we were processing millions of transactions monthly, it was a significant component in the company's budget. I helped build a fraud classifier that when confident would approve a transaction automatically and when not confident would send to this external system. When I joined the project, it was approving 2% of the transactions, when I left it was approving 70% of the transactions, which resulted in significant savings. The improvements mostly came from feature engineering, business understanding (I chose to be embedded in the fraud team) and better modeling.
https://www.reddit.com/user/tabacof