Notes
Notes - notes.io |
close all;
Fs = 1e3; % czestotliwosc probkowania
t = 0:1/Fs:1-1/Fs;
comp1 = cos(2*pi*200*t).*(t>0.7);
comp2 = cos(2*pi*60*t).*(t>=0.1 & t<0.3);
trend = sin(2*pi*1/2*t);
rng default;
Noise = 0.4*randn(size(t));
signal = comp1 + comp2 + trend + Noise;
plot(signal);
pom_co = 0.001; % czas pomiedzy probkami
Fs = 1/pom_co; % czestotliwosc probkowania
t = 0:1/Fs:1-1/Fs;
N = length(signal);
ax = 1 : N; ax = ax * pom_co;
f = (1:N)/(N*pom_co);
set(gcf, 'position',[20,20,400,800])
subplot(6,1,1); plot(ax,signal); xlabel('Seconds');
subplot(6,1,2); histogram(signal);
signal = normalize(signal);
%autokorelacja
[au, la] = autocorr(signal, 200);
subplot(6,1,3); plot(la, au); xlabel('Lag');
%fft
ff = fft(signal);
zoom = 15;
subplot(6,1,4); plot(f(1:int16(N/(2*zoom))), 2*abs(ff(1:int16(N/(2*zoom))))); xlabel("Hz");
%falki
sig = struct('val', signal, 'wavelet', {'morl1'}, 'period', pom_co);
cwtS1 = cwtft(sig);
scales = cwtS1.scales;
MorletFourierFactor = 4*pi / (6 + sqrt (2+6^2));
freq = 1 ./ (scales.*MorletFourierFactor);
subplot(6,1,5); contour(ax, freq, real(cwtS1.cfs));
xlabel('seconds'); ylabel('Hz'); axis([0 ax(end) 0 Fs/(2*zoom)]);
%fractal
[dhseoul, hseoul, cpseoul] = dwtleader(signal);
subplot(6,1,6); plot(hseoul, dhseoul, 'g--*');
xlabel('h'); ylabel('D(h)'); grid on;
%dekompozycja falkowa
lev = 6; %poziom dekompozycji
mra = modwtmra(modwt(signal, "db6", lev));
figure;
set(gcf, 'position', [650,50,500,900]);
subplot(lev+2,1,1);plot(ax,signal); title("Dane");
subplot(lev+2,1,2);plot(ax(200:N),mra(1,200:N)); title("D1")
for kk = 2:lev+1
subplot(lev+2,1,kk); plot(ax(200:N),mra(kk-1,200:N)); title("D"+(kk-1));
end
subplot(lev+2,1,kk+1);
plot(ax, mra(lev+1,:)); title("A"+(kk-1));
xlabel("Time (s)");
figure;
set(gcf, 'position', [1300,50,500,900]);
subplot(lev+2,1,1);plot(ax,signal); title("Dane");
for kk = 2:lev+1
subplot(lev+2,1,kk);
ff = fft(mra(kk-1,:));
zoom = 1*((kk-1));
plot(f(1:int16(N/(2*zoom))), 2*abs(ff(1:int16(N/(2*zoom))))/N);
xlabel("Hz"); title("D"+(kk-1));
end
subplot(lev+2,1,lev+2);
ff = fft(mra(lev+1,:));
zoom = 1*((lev+1));
plot(f(1:int16(N/(2*zoom))), 2*abs(ff(1:int16(N/(2*zoom))))/N);
xlabel("Hz"); title("A"+(kk-1));
%moc dekompozycji falkowej
[c,l]=wavedec(signal,lev,"db6");
[Ea,Ed] = wenergy(c,l);
bar([Ed,Ea]);
%atraktor
dim = 3;
figure;
subplot(2,1,1);
[~,lag] = phaseSpaceReconstruction(signal,[],dim);
lag = 33;
eps = (0.1*((max(signal)) - (min(signal))))^2;
[t, wyk, RQA] = recurent(signal, lag, dim, eps);
RQA_M(poz,:) = RQA(:);
poz = poz + 1;
eRange = 100;
lap = lyapunovExponent(signal,FS,lag,dim,'ExpansionRange',eRange);
plot3(t(:,1),t(:,2),t(:,3)):title("Dane lag = "+lag+"Lap = "+lap);
for ky = 0:500
CCC(ky+1) = nowa_e_d(wyk, 2, ky);
end
%figure
subplot(2,1,2);
spy(wyk);view([-90,90]);
title("Dane"+"RR= "+ RQA(1) + " DET= "+ RQA(2) + " Lmax= "+ RQA(3) + " ENT= " + RQA(5)+ " LAM= " + RQA(6)+ " ENT1= " + RQA(7)+ " VMAX= "+ RQA(8)+ " ent_n= " +RQA(9));
set(gcf,'position',[1800,10,600,1200]);
for kk = 2:lev+1
figure;
subplot(2,1,1);
[~,lag] = phaseSpaceReconstruction(mra(kk-1,200:N),[],dim);
ddane = normalize(mra(kk-1, 200:N));
eps = (0.02*(max(ddane) - min(ddane)))^2;
[t, wyk, RQA] = recurent(ddane, lag, dim, eps/(4*kk));
RQA_M(poz,:)=RQA(:);
poz = poz + 1;
eRange = 100;
lap = lyapunovExponent(ddane,Fs,lag,dim,'ExpansionRange',eRange);
plot3(t(:,1),t(:,2),t(:,3));title("D" + (kk-1) + " lag = " + lag + " Lap = " + lap);
%figure;
subplot(2,1,2);
spy(wyk);view([-90 90]);title("D" + (kk-1)+" RR= "+ RQA(1) + " DET= "+ RQA(2) + " Lmax= "+ RQA(3) + " ENT= " + RQA(5)+ " LAM= " + RQA(6)+ " ENT1= " + RQA(7)+ " VMAX= " + RQA(8)+ " ent_n= " + RQA(9));
set(gcf,'position',[1800,10,600,1200]);
end
figure;
subplot(2,1,1);
[~,lag] = phaseSpaceReconstruction(mra(kk-1,:),[],dim);
eps = (0.02*((max(signal)) - (min(signal))))^2;
[t, wyk, RQA] = recurent(mra(lev+2-1,:), lag, dim,eps);
RQA_M(poz,:) = RQA(:);
poz = poz + 1;
eRange = 100;
lap = lyapunovExponent(mra(kk-1,:),Fs,lag,dim,'ExpansionRange',eRange);
plot3(t(:,1),t(:,2),t(:,3));title("A"+(kk-1)+ " lag = " + lag + " Lap = " + lap);
%figure;
subplot(2,1,2);
spy(wyk);view([-90 90]);title("A"+(kk-1)+" RR= "+ RQA(1) + " DET= "+ RQA(2) + " Lmax= "+ RQA(3) + " ENT= " + RQA(5)+ " LAM= " + RQA(6)+ " ENT1= " + RQA(7)+ " VMAX= " + RQA(8)+ " ent_n= " + RQA(9));
set(gcf,'position',[1800,10,600,1200]);
% atraktor2 ____________________________________________
dim = 3;
figure;
subplot(2,1,1);
[~,lag] = phaseSpaceReconstruction(signal,[],dim);
lag = 33;
[t, wyk1, RQA1] = recurent2(signal, lag, dim);
RQA_M(poz,:) = RQA1(:);
poz = poz + 1;
eRange = 100;
%lap = lyapunovExponent(signal,Fs,lag,dim,'ExpansionRange',eRange);
lap = 0;
plot3(t(:,1),t(:,2),t(:,3));title("Dane lag = "+lag+" Lap = "+lap);
%figure;
subplot(2,1,2);
spy(wyk1);view([-90 90]);title("Dane"+" RR= "+ RQA1(1) + " DET= "+ RQA1(2) + " Lmax= "+ RQA1(3) + " ENT= " + RQA1(5)+ " LAM= " + RQA1(6));
%txt = {'Plotted Data:','y = sin(x)'};
%text(400,500,txt);
set(gcf,'position',[1800,10,600,1200]);
for ky = 0:500
CCC1(ky+1) = nowa_e_d(wyk1, 2, ky);
end
% entropia ____________________________________________
figure
CMSE = ent_w(signal);
plot(CMSE);
CI = trapz(CMSE); %pole pod wykresem CMSE liczone metodą trapezów
% Falki jeszcze raz ____________________________________________
figure
zoom = 15;
contour(ax, freq, real(cwtS1.cfs));
xlabel('Seconds'); ylabel('Hz'); axis([0 ax(end) 0 Fs/(2*zoom)]);
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function CMSE = ent_w(signal)
szereg=signal';
for factor=1:50 %liczba skali czasowych
% tak dobrać factor, aby size(y,2)/factor >=300
p=1; x=1;
while p<(size(szereg,1)-factor);
A(x, factor)=sum(szereg(p:(p+factor-1)))/factor;
p=p+1; %k+factor, wtedy MSE; k=k+1 dla CMSE
x=x+1;
B(factor)=x-1; %tablica tymczasowa przechowująca długości uśrednionych szeregów
end
C(factor)=x-1; %C przechowaj długości uśrednionych szeregów
%SD(j - wiersz - nr ramki; factor - kolumna - poziom uśrednienia)
SD(factor)=std(A(1:(C(factor)), factor));
r(factor)=0.10*SD(factor); %0.2*SD(factor);
end
for factor=1:50
nd=0; nn=0;
for i=1:(C(factor)-2)
for p=(i+1):(C(factor)-2)
if (abs(A(i,factor)-A(p,factor)) < r(factor)) & (abs(A(i+1,factor)-A(p+1,factor))< r(factor))
nn = nn + 1;
if (abs(A(i+2,factor)-A(p+2,factor))< r(factor))
nd = nd + 1;
end
end
end
end
CMSE(factor)=-log2(nd/nn);
end
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [t, wyk, RQA] = recurent(signal, tau, dim, eps)
N = length(signal);
t = zeros(N - tau * dim, dim);
for j = 1 : dim
for i = 1 : (N - tau * dim)
t(i,j) = signal( i + (j - 1) * tau);
end
end
wyk = zeros(N - tau * dim, N - tau * dim);
% wykres rekurencyjny _______________________________
for i = 1:(N-tau*dim)
x = t(i,:);
for j = 1:(N-tau*dim)
x1 = t(j,:);
d = (x-x1).^2;
s = sum(d,"all");
if s < eps
wyk(i,j)=1;
end
end
end
RR = sum(wyk,"All")/((N-tau*dim)*(N-tau*dim));
L = zeros(length(wyk(:,1)),length(wyk(:,1)));
for i = 1:(N-tau*dim)
for j = 1:((N-tau*dim)-i)
if wyk(i+j-1,j) == 1
L(i,j)=1;
end
end
end
dL = zeros(1,length(wyk(:,1)));
for i = 1:length(wyk(:,1))
count = 0;
for j = 1 : length(wyk(1,:)) - i
if L(i,j) == 1
count = count + 1;
end
if L(i,j) == 0
if count > 0
dL(count) = dL(count) + 1;
count = 0;
end
end
end
end
dL = 2* dL;
sdl = 0;
for i = 2:length(dL)
sdl = sdl + dL(i)*i;
end
DET = sdl/(RR*((N-tau*dim)*(N-tau*dim)));
LS = dL(1);
Lmax = 1;
for i = length(dL):-1:1
if dL(i) > 0
Lmax = i;
break;
end
end
DIV = 1/Lmax;
pp(1) = 0;
for i = 1:length(dL)
pp(i) = dL(i)*i;
end
pp = pp./sdl;
ENT = 0;
for j = 2:length(wyk(1,:))
if pp(j) > 0
ENT = ENT + pp(j)*log(pp(j));
end
end
ENT = -ENT;
RQA(1) = RR;
RQA(2) = DET;
RQA(3) = Lmax;
RQA(4) = DIV;
RQA(5) = ENT;
vL = zeros(1,length(wyk(:,1)));
count = 0;
for i = 1:length(wyk(:,1))
count = 0;
for j = (i+1):length(wyk(:,1))
if wyk(i,j) == 1
count = count + 1;
end
if ((wyk(i,j) == 0) | (j == length(wyk(:,1))))
if count > 0
vL(count) = vL(count) + 1;
count = 0;
end
end
end
end
sdl = 0;
for i = 2:length(vL)
sdl = sdl + vL(i)*i;
end
LAM = sdl/(RR*((N-tau*dim)*(N-tau*dim)));
pp1(1) = 0;
for i = 1:length(vL)
pp1(i) = vL(i)*i;
end
pp1 = pp1./sdl;
ENT1 = 0;
for j = 2:length(wyk(1,:))
if pp1(j) > 0
ENT1 = ENT1 + pp1(j)*log(pp1(j));
end
end
ENT1 = -ENT1;
RQA(6) = LAM;
RQA(7) = ENT1;
LS = vL(1);
Vmax = 1;
for i = length(vL):-1:1
if vL(i) > 0
Vmax = i;
break;
end
end
RQA(8) = Vmax;
RQA(9) = nowa_e(wyk, 1000, 2);
RQA(10) = nowa_e_d(wyk, 2, 0);
end
|
Notes.io is a web-based application for taking notes. You can take your notes and share with others people. If you like taking long notes, notes.io is designed for you. To date, over 8,000,000,000 notes created and continuing...
With notes.io;
- * You can take a note from anywhere and any device with internet connection.
- * You can share the notes in social platforms (YouTube, Facebook, Twitter, instagram etc.).
- * You can quickly share your contents without website, blog and e-mail.
- * You don't need to create any Account to share a note. As you wish you can use quick, easy and best shortened notes with sms, websites, e-mail, or messaging services (WhatsApp, iMessage, Telegram, Signal).
- * Notes.io has fabulous infrastructure design for a short link and allows you to share the note as an easy and understandable link.
Fast: Notes.io is built for speed and performance. You can take a notes quickly and browse your archive.
Easy: Notes.io doesn’t require installation. Just write and share note!
Short: Notes.io’s url just 8 character. You’ll get shorten link of your note when you want to share. (Ex: notes.io/q )
Free: Notes.io works for 12 years and has been free since the day it was started.
You immediately create your first note and start sharing with the ones you wish. If you want to contact us, you can use the following communication channels;
Email: [email protected]
Twitter: http://twitter.com/notesio
Instagram: http://instagram.com/notes.io
Facebook: http://facebook.com/notesio
Regards;
Notes.io Team