Confusion matrix outputs NaN values after classification using Neural Network

11 次查看（过去 30 天）
Santa Maria
Santa Maria 2016-12-6
回答： Greg Heath 2016-12-6
I am trying to classify data containing 10 classes using neural network. My data is in the form of wavfiles. I computed my feature vector using n-point FFT. The problem is that after classification, I am getting NaN as an output from the confusion matrix, although my input to the Neural Network doesnt contain any NaN values. When I classify the data for 3 classes, there is no such problem. But if I increase the number of my classes from 5 and more, I get NaN values.

请先登录，再回答此问题。

回答（1 个）

Greg Heath
Greg Heath 2016-12-6
Isn't it obvious?
1. Rows 9 and 10 contain nothing but zeros.
2. Ratios and percentages obtained by dividing by zero will yield NaNs
Hope this helps.
Thank you for formally accepting my answer
Greg

类别

在 Help Center 和 File Exchange 中查找有关 Deep Learning Toolbox 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by