how to predict next value using time series data?

I am having 2 patient glucose data(y(t)) and the inputs feature values (time,meal) for 3 days. How can i predict the next time step glucose value. I am little bit confuse about how to give data to neural network. input= time, meal output=glucose data depending on time and meal, output glucose level is calculated. So how should i store my data in excel sheet so that i can do the good prediction.

[Greg Heath]

Do you have enough data??

 [ I N ] = size(input) % = ?
 [ O N ] = size(target) % = ?

Greg

[madhuri panwar]

Thank You for your response sir. size(input1)= 144 data(3 days ,data taken every 30 min) size(input2)=144 data(3 days ,data taken every 30 min) size(output)=144 data

[Greg Heath]

I do not understand . The command size NEVER yields a single number.

What are the two-dimensional sizes of your input and target matrices? The 2nd, N, is the number of input-output pairs.

[madhuri panwar]

Dear Sir, Thank You once again for responding me.

[I N(1)]= Size(input1)=144 samples [I N(2)] =Size(input2)=144 samples [I N(3)]= Size(input3)=144 samples [O N] = Size (target)=144 samples samples are taken for three day and 1 sample in every 30 min (48 for 24 hours) (48*3days=144)

Here i am having 3 inputs- time, meal and insulin output - glucose

[Greg Heath]

Why do you try to make it so confusing? Just type the following into the command line and tell me what you get

 [ I N ] = size(input)

 [ O N ] = size(target)

Greg

Answer 1

[Greg Heath]

 OK I get it:

 [ I N ] = size(input) % [ 3 144 ]

 [ O N ] = size(target) % [ 1 144]

Hope this helps.

Thank you for formally accepting my answer

Greg