I am currently looking at how performance of automated update of datasets can affect advanced statistical models and ML techniques. I was wondering if cross-entropy could help stating clearer the potential impact on research, models predictions and classification outcome - alongside descriptive statistics.
I am not convinced as I feel some statistical test may be able to be more effective. However, cross-entropy may be a better metric.
I am looking forward to hear whether it is worth the pursue or not. I am not convinced. Thank you for your help.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs
More Patricia Ryser-Welch's questions See All
Similar questions and discussions