# Confidence interval

With

data = {{0.002, 2.51053}, {0.004, 2.54217}, {0.006, 2.55543}, {0.008, 2.54247}};


I obtain a linear fit 2.51038 + 5.454 x; now I want to get the 95% confidence interval from linear regressions. when I do it with Mathematica, it gives me a pair of data set for the CI. I want to add the CI to x -> 0.000717263, but I don't know how to interpret that.

• The size 4 of your data is too small to make reliable estimates, e.g. see researchgate.net/post/… . – user64494 Jul 27 '20 at 6:39
• Pay your attention to the result of glm = GeneralizedLinearModelFit[data, x, x] glm["ParameterTable"]$$\begin{array}{l|llll} \text{} & \text{Estimate} & \text{Standard Error} & \text{z-Statistic} & \text{P-Value} \\ \hline 1 & 2.51038 & 0.019372 & 129.588 & 0. \\ x & 5.454 & 3.53682 & 1.54206 & 0.123058 \\ \end{array}$$ – user64494 Jul 27 '20 at 8:33

Is this what you want?

data = {{0.002, 2.51053}, {0.004, 2.54217}, {0.006, 2.55543}, {0.008, 2.54247}};

lm = NonlinearModelFit[data, a x + b, {a, b}, x]

bands[x_] = lm["MeanPredictionBands", ConfidenceLevel -> .95]

Show[ListPlot[data],
Plot[{lm[x], bands[x]}, {x, 0, 0.01}, Filling -> {2 -> {1}}],
Frame -> True, Axes -> False, PlotRange -> All]


bands[0.000717263]


{2.44077, 2.58781}