"For some data with much the noise, it is usually necessary to smooth the data to reduce the nois .."

Lina RaqForum 43 No.
359 View • 1 Years ago

sg data smoothing(polynomial smoothing)

For some data with much the noise, it is usually necessary to smooth the data to reduce the noise.

For example, a set of noisy data as below, use the sg() function to polynomial smooth the data

	A
1	=file("sgdata.xlsx").xlsimport@w()
2	=A1.(~(1))
3	=A1.(~(2))
4	=sg(A3,1,15).conj()
5	=canvas()
6	=A5.plot("NumericAxis","name":"x")
7	=A5.plot("NumericAxis","name":"y","location":2)
8	=A5.plot("Line","markerStyle":0,"axis1":"x","data1":A2,"axis2":"y", "data2":A3)
9	=A5.plot("Line","markerStyle":0,"lineColor":-65536,"axis1":"x","data1":A2,"axis2":"y", "data2":A4)
10	=A5.draw@p(600,600)

A1 Read the data as a secondary sequence

A2 The sequence number

A3 Original data

A4 Perform 15-point linear smoothing on A3 data (n=1,m=15), and return the smoothed data, as the red line in the figure. In sg(A,n,m,d), taking no derivative when the parameter d is omitted means smoothing only; d=1, indicating the smoothing result of the first derivative, which can eliminate the constant term error; d=2 Second derivative smooth, can eliminate the linear error

A5-A10 Draw the data before and after smoothing and observe the smoothing result. If the result is not good, the value of n, m, d can be adjusted until it is good.