M精編b線性回歸擬合yyMatlab線性回歸(擬合)(x,x,i1i2ipi記x=xx…xx1pxx…xxx…xx…xx，(y1y=，2=，：yn如果i的置信區(qū)間(bint的第i+1行)不包含0，則在顯著水平為a時(shí)拒絕i=0的假設(shè)，認(rèn)為變量xi是顯著的．xysinx值點(diǎn)xx點(diǎn)yxxotxyoxxyy性回歸分析)：beta=nlinfit(x,y,fun,beta0)x據(jù),fun數(shù)形式),beta數(shù)估計(jì)初值,x=lsqcurvefit(fun,x0,xdata,ydata)xdata:自變量數(shù)據(jù),ax合結(jié)果),beta:估計(jì)出的回歸系數(shù),J:Jacobian矩陣,Model:事先用m-文件定義的非線性函數(shù)betax1=[0.5,0.4,0.3,0.2,0.1];x0.5,0.2,0.4,0.6];x.4,1.0,1.4,1.8];y703,0.583,0.571,0.126]’;yab*x2+c*x3+d*(x2.^2)+e*(x3.^2)betaxyybetabeta(2)*x2+beta(3)*x3+beta(4)*(x2.^2)+beta(5)*(x3.^2);x,0.3,0.2,0.1;0.3,0.5,0.2,0.4,0.6;1.8,1.4,1.0,1.4,1.8]';y.703,0.583,0.571,0.126]';beta,1,1]';a塊，記錄了養(yǎng)護(hù)日期(日)及抗壓強(qiáng)度y(kg/cm2)的數(shù)據(jù)：x7212856];m=0.1083rrand(1,12)-0.5;y687376828699];myfuncinlinebeta)+beta(2)*exp(beta(4)*x)+beta(3)*exp(-betaxbeta','x');betanlinfitxy,myfunc,[0.50.50.50.5]);abetakbetakbeta,m=beta(4)tthemodelxmaxxyakexpmxxkexpmxxplotxyoxxyyr)k9k1.2lsqnonlin非線性最小二乘(非線性數(shù)據(jù)擬合)的標(biāo)準(zhǔn)形式為「f(x)]F(x)=|fx)|設(shè)Lfm(x)」則目標(biāo)函數(shù)可表達(dá)為ximinF(x)=fi則目標(biāo)函數(shù)可表達(dá)為xix=lsqnonlin(fun,x0,lb,ub,options)%options為指定優(yōu)化參數(shù)，若x沒有界，則lb=[]，ub=[]。xresnorm]=lsqnonli()%resnorm=sum(fun(x).^，2)即解x處目標(biāo)函數(shù)值。xfunctionF=myfun(x)k:10;F=2+2*k-exp(k*x(1))-exp(k*x(2));terminatedsuccessfullyrmofthecurrentstepislessthanOPTIONSTolXorm數(shù)關(guān)系為ydata=F(x,xdata)，但不知道系數(shù)向量x。今進(jìn)行曲線擬合，求x使得下式成xdataydatax=lsqcurvefit(fun,x0,xdata,ydata,lb,ub)qcurvefitfunxxdataydatalbuboptionsbfun為擬合函數(shù)，其定義方式為：x=lsqcurvefit(@myfun,x0,xdata,yd，at)中myfun已定義為functionF=myfun(x,xdata)F=…%計(jì)算x處擬合函數(shù)值fun的用法與前面相同；resnorm=sum((fun(x,xdata)-ydata)，.^)在x處殘差的平方和；residual=fun(x,xdata)-yd，at在x處的殘差；curvefitmyfunxxdataydataterminatedsuccessfullyctionvaluechangingbylessthanOPTIONSTolFun21etanlinfitXyfunbetasx始猜測值xesnormthevalueofthesquarednormoftheresidualatxsumfu