How can I use deep learning network to extract features from an RSSI samples dataset? My dataset contains nearly 45000 samples.

I am trying Deep Learning Neural Network (DNN to extract features embedded within a wireless received signal strength (RSS) dataset consisting of 45K samples. How do I feed the network with this dataset, and what is the actual size of the dataset for DNN in MATLAB?

Answer 1

[Srivardhan Gadila]

Features learned by the deep neural network depends on the input and the target data i.e., for what kind of problem is your network defined and trained e.g: Classification, Regression etc.

There is no defined size for the dataset, it completely depends on the problem you are training the network for, how different are the data samples in your dataset etc.

You may refer to some feature extraction examples in Image, audio domains to start with

1. https://mathworks.com/discovery/feature-extraction.html
2. https://mathworks.com/help/audio/feature-extraction-and-deep-learning.html