train a Neural network with different input output sizes

Hi, I would like to train a Neural Network (pattern-recognition) where I have vectors of 500 points (input) each belong to single output value.

The input vectors of 500 points are a timeseries, contains the Info about a sharp peak on a certain posistion which belongs a single Value.

So how to tran a network where I have hundrets of

Inputs (1x500) and outputs (1x1).

Thx,

Chris

Answer 1

[Matt J]

If it's a sequence-to-one regression then this syntax of trainNetwork would be appropriate,

net = trainNetwork(sequences,responses,layers,options)