User`s guide

predint

4-124

If simopt is off, nonsimultaneous bounds are calculated. If simopt is on,

simultaneous bounds are calculated. Nonsimultaneous bounds take into

account only individual

x values. Simultaneous bounds take into account all x

values.

[ci,ypred] = predint(...) returns the predicted (fitted) value of fresult

evaluated at

Example Generate some data and add noise.

x = (0:0.2:10)';

coef = [2 -0.2];

rand('state',0)

y = coef(1)*exp(coef(2)*x) + (rand(size(x))-0.5)*0.5;

Fit the data using a single-term exponential and define the range over which

prediction bounds are calculated.

fresult = fit(x,y,'exp1');

Return the prediction bounds for the function as well as the predicted values of

the fit using nonsimultaneous and simultaneous bounds with a 95% confidence

level. For nonsimultaneous bounds, given a single predetermined predictor

value, you have 95% confidence that the true function lies between the

confidence bounds. For simultaneous bounds, you have 95% confidence that the

function at all predictor values lies between the bounds.

[c1,ypred1] = predint(fresult,x,0.95,'fun','off');

[c2,ypred2] = predint(fresult,x,0.95,'fun','on');

Return the prediction bounds for new observations as well as the predicted

values of the fit using nonsimultaneous and simultaneous bounds with a 95%

confidence level. For nonsimultaneous bounds, given a single predictor value,

you have 95% confidence that a new observation lies between the confidence

bounds. For simultaneous bounds, regardless of the predictor value, you have

95% confidence that a new observation lies between the bounds.

[c3,ypred3] = predint(fresult,x,0.95,'obs','off');

[c4,ypred4] = predint(fresult,x,0.95,'obs','on');