First, Data transformation technology
In order to ensure the quality of the modeling and the accuracy of the results of the system analysis, the original data should be dimensionally processed.
1. Definition
With a sequence, it is mapped into a data transformation of sequence x to sequence Y.
(1) F is an initialization transformation.
(2) F is the mean value transformation.
(3) F is a percent transform
(4) F is a multiplier transformation
(5) F is the normalized transformation of one of the x0>0 values
(6) F is the extreme difference maximum value transformation
(6) F is the interval value transformation
Second, relevance analysis
1. Definition
The reference sequence is compared (subject) and the comparison number is the comparison value (not the subject)
But because each moment has a resolution factor, too scattered, so define an indicator
2. Demo
MATLAB solution
function [y] = relevancy (Refer,compare,p,rank)%refer Reference series (row vector), compare comparison sequence%p is the resolution factor, the default is 0.5%rank for the column vector, The reference series and the comparison sequence are correlated or unrelated when the%rank is 1 to indicate that the same increment is correlated (default), rank 0 indicates that the same increment is irrelevant%y returns a column vector that reflects the correlation degree [a,b]=size (compare); if (nargin <3) P=0.5;endif (nargin<4) rank=ones (a,1);end% Reference Series, Comparison of sequence initialization refer=refer/refer (1);For i=1:aif (rank (i)==1)compare (i,:)=compare (i,:)/compare (i,1); Else Compare (i,:)=compare (i,1)./compare (i,:); endend% Calculate the difference for I=1:aFor J=1:bdiff (I,j)=compare (I,j)-refer (j); endend% Calculating relevancy diff=abs (diff);For i=1:asum= 0; For J=1:bMax1=max (Max (diff)); min1=min (Min (diff ')); sum=sum+ (MIN1+P*MAX1)/(diff (i,j) +p*max1),% correlation degree calculation FormulaEnd y (i,1)=sum/b;% correlation degree averagingEnd
Modeling Algorithm (10)--correlation degree analysis of grey theory