How do I pre-process and process non-image data for Deeplearning?

I am trying to make predictions using deep learning and CNN on non-image data but running into difficulty processing csv file for training.

[Image Analyst]

What kind of data do you have? What is it's size? Can you reshape it into an image of the proper size, even if you have to have some regions zero (which the training will learn to ignore)?

[Mye'on Williams]

Its is a 10000x14 table which has values such as:

2.9 | 3.0 | 8.3 | 9.7 | 3.7 | -0.7 | -1.25 | -1.7 | 0.6 | 0.8 | 0.8 | 0.9 | 0.5 | 'unstable'

Data can not be reshaped to an image. I am tasked to use deep learning networks to a non-image dataset. Will CNN on non-image binary predictions be the way to go?

[yanqi liu]

yes，sir，may be upload your data file，use reshape to get 4-D input，and train cnn

Answer 1

[Prateek Rai]

Hi,

You can refer to "Signal Processing Applications" section of the following Preprocess Data for Domain-Specific Deep Learning Applications MathWorks Documentation page to know more about pre-processing non-image data for deep learning.