smoothing of an array of data

I have to smooth a set of data(a array of data).I would like to implement savitzky golay smoothing filter but without the use of built in function.Can any1 please help me out??Moreover after the smoothing is done how can i classify the states of the data using intelligent pattern recognition??

[Jan]

What is the reason of now using builtin functions? It is absolutely impossible to avoid *all* builtin functions, e.g. plus, minus, eq, filter etc.

[Durga Lakshmi]

i know that it reduces complexity but i need to know and understand the working of the slavitzky golay filter.So asked for the entire code.

Answer 1

[Jan]

This is A Savitzky-Golay filter without dependencies to the Signal Processing Toolbox: FEX: fSGolayFilt. It is remarkably faster than the built-in method and accepts SINGLE values.

Please post more details about the pattern recognition problem. There are millions of different patterns and thousands of methods to recognize them more or less accurate.

[Durga Lakshmi]

thanx fr answering! I am using a sgolay function

smtlb = sgolayfilt(mtlb,3,41); % Applying 3rd-order filter

Can u tell me why are we using 3 rd order filter and not any other?If we can use other orders then which order should i choose.

My array consist of industrial data which has data of several states of operation.I need to classify the states using pattern recognition.I am open for any method but it must be simple and concise.