Should I use deep learning to solve my task?

I need to predict the performance (CPI cycles-per-instruction) of 90 machines for the next hour (or day). Each machine has a thousand records (e.g. CPU and memory usage).

Currently, I am using a neural network with one hidden layer for this task. I have 9 inputs (features), 23 neurons in the hidden layer, and one output. I am using the Levenberg-Marquardt algorithm. Examples of the inputs (or features) are the CPU and memory capacity and usage, and the machine_id. and Output is performance. I have 90 machines. Currently, I get an MSE of $$0.1$$ and an R of $$0.80$$.

My dataset consist of 30 days. I trained my network for the first 29 days, and I use day 30 to test.

I have been advised to use deep learning to have more flexibility and improve the MSE and R results. Could deep learning be helpful in this case? If yes, which deep learning model could I use to improve the results?

• Hi Jou. It's unclear what you're asking. What do you want to achieve? What have you tried? – nbro Dec 9 '19 at 23:30
• You should explain the problem you're trying to solve. What do you want to predict? What results are you currently obtaining with the MLP? – nbro Dec 9 '19 at 23:42
• Take a look @nbro. and I appreciate your time – jou Dec 9 '19 at 23:51
• You say you want to predict the "machine performance". But what do you mean by "machine performance" exactly? You say you have 90 machines. Do you mean that you have a training dataset of 90 observations one for each of these machines? Or you want to predict the performance of 90 machines, which would then be the test cases of your trained model? – nbro Dec 11 '19 at 14:14
• No. Deep Learning efficient for big networks and samples of extremely high dimensions. It don't make sense to use it for 9 element vectors as samples. – mirror2image Dec 12 '19 at 7:10