classification of signls using Matlab

which characteristic features can i use to classify between signals of different shapes. I have signals of different shapes: some looks more like sinusoidal, and some looks like triangular signals, which mathematical parameter could be best used to describe the shape of the signal ?? my signals all are sample at the same frequency which is 30 Hz.

what is the best parameters, which could be calculated to describe the shape of the signal ?!??

Answer 1

[Image Analyst]

Depends on what you want to characterize it. Maybe the mean will do what you need. Maybe you need the mean and the standard deviation. Maybe you need several moments. Who knows how many numbers you need to describe the signal. In the end, all the elements of the signal describe it perfectly, accurately, and exactly. Maybe you want the spectrum by taking the fft. It really depends on what you want to do and why you want to do this.

[hamzah Hmeidi]

I am looking at medical signals which i could use to classify between groups. I have noticed that the different groups have different signal shapes. so i was thinking about a way to classify between those groups mathematically. the signals are 300 seconds each sampled at 30 Hz, so i was thinking to calculate the RMS value to describe them. I don't know if it is a good idea!! is there is another useful parameter which can describe their shapes, knowing that some of them are more like a sinusoidal signals for one group, and for the other group they are more toward triangular signals.

[hamzah Hmeidi]

so i don't have the function of the signals, i have only the samples of each signal which i used to plot the signal using Matlab, and i have noticed the shape difference among the groups.