Error using Regression​SVM.prepar​eData. anybody can help me

Here I give my coding and i have error on this:

Error in newavelet (line 49) mdl = fitrsvm(xt,yt,'KernelFunction','linear','KernelScale','auto');

close all; clear variables; clc;

DT=xlsread('imf9'); N=length(DT); input_lag=2;%no. of input IN=input_lag; dt=log(DT);

%matrix lag yg digunakan utk regression XLAG = lagmatrix(dt,[0:IN]); %change matrix if input_lag change N1=length(XLAG)-IN; %length utk panjang N without NA observation F=61; %No. of forecast N2=N1-F; %No. of training

%training - use (IN+1) due to after lagmatrix N/A data for no. of inputs xt=XLAG(IN+1:N-F,2:IN+1); yt=XLAG(IN+1:N-F,1);

%validation - use (N_ORI-1) due to after log return data reduce 1 xv=XLAG(N-F+1:N,2:IN+1); yv=XLAG(N-F+1:N,1);

%original data for calculating error ytOri=DT(IN+1:N-F);%training data yvOri=DT(N-F+1:N);%validation data

%define matrix utk regression a=ones(N2,1);%training data b=ones(F,1);%validation data X=[a xt];%training data x=[b xv];%validation data

d1='db2'; Dc=3;%wavelet decomposition level

[w, w1]=waveletdwt(IN,Dc,d1,xt,xv); Z=[a w];%wavelet data utk training Z1=[b w1];%wavelet data utk validation

%xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx % SVM %xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

mdl = fitrsvm(xt,yt,'KernelFunction','linear','KernelScale','auto'); yts = predict(mdl,xt); yvs = predict(mdl,xv);

%transform log tourism data to tourism data transformYts=exp(yts); transformYvs=exp(yvs);

YTS=transformYts; YVS=transformYvs; M1(1)=sqrt(mse(YTS-ytOri)); M1(2)=sqrt(mse(YVS-yvOri)); M1(3)=mae(YTS-ytOri); M1(4)=mae(YVS-yvOri); M1(5)=corr(YTS,ytOri); M1(6)=corr(YVS,yvOri); %M11(5)=1-sum((y1-yt).^2)/sum((y1-mean(y1)).^2); %M11(6)=1-sum((pd2-fr).^2)/sum((fr-mean(fr)).^2);

%xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx % Wavelet-SVM %xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

wmdl = fitrsvm(Z,yt,'KernelFunction','linear','KernelScale','auto'); Wyts = predict(wmdl,Z); Wyvs = predict(wmdl,Z1);

%transform log tourism data to tourism data WYTS=exp(Wyts); WYVS=exp(Wyvs);

        M11(1)=sqrt(mse(WYTS-ytOri));
        M11(2)=sqrt(mse(WYVS-yvOri));
        M11(3)=mae(WYTS-ytOri);
        M11(4)=mae(WYVS-yvOri);
        M11(5)=corr(WYTS,ytOri);
        M11(6)=corr(WYVS,yvOri);
        %M11(5)=1-sum((y1-yt).^2)/sum((y1-mean(y1)).^2);
        %M11(6)=1-sum((pd2-fr).^2)/sum((fr-mean(fr)).^2);

%xxxxxxxxxxxxxxxxxxxxxxxxxxx % result & plot graph %xxxxxxxxxxxxxxxxxxxxxxxxxxx

digits(4) SVMR=sym(M1,'d') WSVMR=sym(M11,'d') g1=1:F; plot(g1,yvOri,'r',g1,WYVS,'g') % plot original data vs result from wavelet svm