表九：旅游人数预测值与真实值比较（百万人）
年份 预测值 真实值 百分绝对误差 0.70.650.60.550.50.450.40.35原始曲线神经网络拟合曲线2006 1145.8 1394 17.80% 2007 1576.2 1610 2.10% 2008 1740.2 1712 1.65% 2009 1643.1 1902 13.61% 0.10.20.30.40.50.60.70.80.91
图3：旅游人数的真实值与拟合值
四、本底趋势线模型 旅游总收入：
function f=curvefunl(x,t)
f=x(1)*t+x(2)+x(3)*sin(x(4)*t+x(5))
t=1:15;
c=[1375.7 1638.38 2112.7 2391.18 2831.92 3175.54 3522.37 3878.36... 4567.14 4710.71 5285.86 6229.7 7770.6 8749.295918 10183.7]'; x0=[1,1,1,0.1,0.1];
x=lsqcurvefit('curvefunl',x0,t,c); f=curvefunl(x,t); x
结果： x =
557.6560 101.9953 84.8500 31.9949 5.5929
t=1:15;
24
c=[1375.7 1638.38 2112.7 2391.18 2831.92 3175.54 3522.37 3878.36... 4567.14 4710.71 5285.86 6229.7 7770.6 8749.295918 10183.7]'; x=[557.6560 101.9953 84.8500 31.9949 5.5929]; f=x(1)*t+x(2)+x(3)*sin(x(4)*t+x(5)) plot(t,c,'k*',t,f,'r')
120001000080006000400020000051015
旅游游客人数 x=1:15; y=[629 639.5 644 695 719 744 784 878 1047.46 1102 1212 1394 1610 1712 1902]; a=polyfit(x,y,2) z=polyval(a,x);
plot(x,y,'k*',x,z,'r')
结果： a =
7.6967 -32.3245 669.7975
25
200018001600140012001000800600051015
26