Hello, dear researchers
How can I improve the (Matthews correlation coefficient)MCC parameter? That is, the results are related to a deep network.
The target is defined as a binary prediction (0 or 1).
Please write your guide clearly and introduce the appropriate reference
Thank you for your guidance
The MCC is measured on how well your model fits your data. I guess you need better predictors, or am I missing something?
Hi, dear Ronán Michael Conroy
thank you for your advice
Is there an operation that can be used to improve MCC, such as averaging
I guess the message is that your network is, as we say, a not-work.
What happens when you use a simple linear model like logistic regression? Networks rarely outperform logistic to a meaningful degree, and if a logistic doesn't perform well I would take it that the network isn't going to work out.
- Badges
- Science topic
- Similar topics
- Comment