Your question is Handling Collinearity in High-Frequency Data. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How do you handle highly collinear features in a high-frequency financial dataset without losing predictive power?