Is there any way to increase accuracy with this Convolution Neural Network?

1 view (last 30 days)
Chris Basic
Chris Basic on 2 Apr 2020
Edited: Prateek Rai on 8 Oct 2021
I have a convolution neural network with layers like this:
layers = [ ...
imageInputLayer([1 5 1])
convolution2dLayer(1,10)
fullyConnectedLayer(2)
softmaxLayer
classificationLayer];
The idea is that I have multiple imputs of one row by 5 columns, a convolution 2d layer, and I am trying to classify this set as either 70 or 83 from a set of numbers. I usually get an accuracy of 50-55 percent. Is there anything that can be done to increase the accuracy or is it not suitable to use a CNN to classify from a random set of numbers?

Answers (1)

Prateek Rai
Prateek Rai on 8 Oct 2021
Edited: Prateek Rai on 8 Oct 2021
To my understanding, you have multiple inputs of one row by 5 columns and you want to classify this set as either 70 or 83 from a set of numbers.
Since you are not getting satisfactory results, you can use "Classification Learner App" and decide which algorithm to use. You can Interactively train, validate, and tune classification models with "Classification Learner App". This will help you find the best model for your data.
You also refer to Classification Learner App MathWorks documentation page to learn more on Classification Learner App. To help you decide which algorithm to use, you can also refer to Train Classification Models in Classification Learner App.

Products


Release

R2018b

Community Treasure Hunt

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

Start Hunting!