A note on variability of interval data (Q626264)

One observes intervals \([a_i,b_i]\), \(i=1,\dots,n\) to which the unobserved data points \(x_i\in[a_i,b_i]\) belong. The aim is to calculate strict upper and lower bounds for the possible sample variance and the mean absolute deviation of \(x_i\). The authors discuss quadratic and linear programming algorithms for the solution of this problem. An iterative algorithm is proposed for the minimization of the variance, which is based on the idea to minimize \(\sum_i(x_i-p)^2\) at each step for fixed \(p\) and to take for the next step \(p=\sum_i x_i/n\). An analogous idea can be used for maximization also. A real life data set example is presented.