How do I get rid of noise in my data matrix?

I have a set of data that cycles from low x&y values to high x&y values. Using the ismember function, I've been able to find the max/min values for each cycle. However, becuase this is actual data, there is a lot of noise in the data. So instead of getting one max value and one min value for every cycle, I am getting multiple min/max values. I know the index values for the points and I know each value should be spaced out by about 500 data points. So if I see that the index values between two points are less than 450, I know that there are multiple min/max values for that cycle. How could I write a code that automatically detects and eliminates those noisy data points from my data matrix? I'm guessing it has something to do with an if-then statement, but I'm not sure how. Thanks!

[Bob Thompson]

Why not just use some kind of root mean square smoothing?

Answer 1

[John D'Errico]

An if statement? Well you can try anything you want. But you won't be happy. Instead, just use a simple moving average filter, or a Savitsky-Golay filter, or a Gaussian blur filter. Any of these tools will smooth your data.