Basic question about Accuracy in EON Tuner

[doyney]

Hi,

Is the idea with the accuracy matrix in the EON tuner to have all 4 boxes be as close to 100% accuracy?
Or is it for Neg-Neg and Pos-Pos (in my example) to be 100% and Neg-Pos and Pos-Neg to be 0%?

I’m pretty new to machine learning, so it’s just a basic question. Any other helpful info would be appreciated, i’m trying to learn as much as I can right now.

[louis]

Hello @doyney ,

Yes Neg-Neg and Pos-Pos should be as close as 100% possible and Neg-Pos and Pos-Neg should be as close 0%.

Here is a great explanation by @shawn_edgeimpulse if you want to sharpen your knowledge:

How to Evaluate a Model - Introduction to Neural Networks | Coursera

Or here just a basic understanding with more labels:

Classification (Keras)

Regards,

Louis