Analysts often seek to compare representations in high-dimensional space, e.g., embedding vectors of the same word across groups. We show that the distance measures calculated in such cases can exhibit considerable statistical bias, that stems from uncertainty in the estimation of the elements of those vectors. This problem applies to Euclidean distance, cosine similarity, and other similar measures. After illustrating the severity of this problem for text-as-data applications, we provide and validate a bias correction for the squared Euclidean distance. This same correction also substantially reduces bias in ordinary Euclidean distance and cosine similarity estimates, but corrections for these measures are not quite unbiased and are (non-intuitively) bimodal when distances are close to zero. The estimators require obtaining the variance of the latent positions. We (will) implement the estimator in free software, and we offer recommendations for related work.

Ideal point discriminant analysis is a classification tool which uses highly intuitive multidimensional scaling procedures. However, in the last paper, Takane wrote about it. He concludes that the interpretation is rather intricate and calls that a weakness of the model. We summarize the conditions that provide an easy interpretation and show that in maximum dimensionality they can be obtained without any loss. For reduced dimensionality, it is conjectured that loss is minor which is examined using several data sets.

Given any two vertices u, v of a random geometric graph G(n, r), denote by dE(u, v) their Euclidean distance and by dE(u, v) their graph distance. The problem of finding upper bounds on dG(u, v) conditional on dE(u, v) that hold asymptotically almost surely has received quite a bit of attention in the literature. In this paper we improve the known upper bounds for values of r=ω(√logn) (that is, for r above the connectivity threshold). Our result also improves the best known estimates on the diameter of random geometric graphs. We also provide a lower bound on dE(u, v) conditional on dE(u, v).

Consider two independent Goldstein-Kac telegraph processes X1(t) and X2(t) on the real line ℝ. The processes Xk(t), k = 1, 2, describe stochastic motions at finite constant velocities c1 > 0 and c2 > 0 that start at the initial time instant t = 0 from the origin of ℝ and are controlled by two independent homogeneous Poisson processes of rates λ1 > 0 and λ2 > 0, respectively. We obtain a closed-form expression for the probability distribution function of the Euclidean distance ρ(t) = |X1(t) - X2(t)|, t > 0, between these processes at an arbitrary time instant t > 0. Some numerical results are also presented.