I am working on a project of " Fraud detection using deep learning" . For that I have a dataset containing some numerical attributes . Now the task is to use CNN for the above purpose. Please guide me how can we do that as I have observed that it can be used only for image classification. I would appreciate if I get some research paper or blogs related to that.

  • $\begingroup$ Your question seems quite wide. Moreover, as you mention, there's no reason to use a CNN rather than some other architecture. Perhaps you should ask your lecturer of TAs for further guidance. $\endgroup$ – Yuval Filmus Oct 8 '18 at 6:58
  • $\begingroup$ Isn't all data in computers basically numbers? $\endgroup$ – David Richerby Oct 8 '18 at 8:56

Convolutional neural networks are just one of many models for classifing data. All models can be used for any data and they differ only in performance.

When you feed an image to the CNN (or any other model), the model does not “see” the image as you see it. It “sees” numbers that describe each pixel of an image and does all calculation using those numbers.
As was pointed in the comments; at the end, everything is effectively a number in the computer.

Your task asks you to use CNN because it will most likely perform best (in comparisson to other models) at this specific task. CNNs are very good at classifying objects that are not too similar, but have some shared features presented in a not too obvious way.

You can read about classification models here and I found a paper about detecting fraud with CNNs that might help you.

| cite | improve this answer | |

Not the answer you're looking for? Browse other questions tagged or ask your own question.