Can I use Convolution Neural Network (CNN) as a two class classifier?

I have two class data sets of 1D signal. I want to classify them using Convolution Neural Network.Is it possible? * Is there any build in functions available?

Answer 1

[Zuber Khan]

Hi,

Yes, you can apply Convolution Neural Network (CNN) to 1-D data as well for classification. CNNs are not limited to 2D images; they can be effectively applied to 1D data such as time series, audio signals, or any sequential data.

You can use 1D convolutional layers (Conv1D), followed by pooling layers (MaxPooling1D), and fully connected layers for the same.

A 1-D convolutional layer applies sliding convolutional filters to 1-D input. In MATLAB, you can create this layer by invoking "convolution1dLayer" method present in Deep Learning Toolbox. For more information, refer to the following documentation:

https://www.mathworks.com/help/deeplearning/ref/nnet.cnn.layer.convolution1dlayer.html

Similarly, a 1-D max pooling layer performs downsampling by dividing the input into 1-D pooling regions, then computing the maximum of each region. To know more about "maxPooling1dLayer" method, please refer to the following documentation:

https://www.mathworks.com/help/deeplearning/ref/nnet.cnn.layer.maxpooling1dlayer.html

You can use both these layers while creating the network architecture. Apart from building network, you would need to preprocess your signal data in a format appropriate for training deep learning networks.

Although, there is no direct resource but you can have a look at the following example as reference which shows how to create and train a simple convolutional neural network for deep learning based classification.

https://www.mathworks.com/help/deeplearning/ug/create-simple-deep-learning-network-for-classification.html

I hope it will answer your query.

Regards,

Zuber