We propose a semiparametric estimator for multinomial discrete choice models. The term “semiparametric” refers to the fact that we do not specify a particular functional form for the error term in the random utility function and we allow for heteroskedasticity and serial correlation. Despite being semiparametric, the rate of convergence of the smoothed maximum score estimator is not affected by...

Expectation-maximization (EM) is an iterative algorithm for finding the maximum likelihood or maximum a posteriori estimate of the parameters of a statistical model with latent variables or when we have missing data. In this work, we view EM in a generalized surrogate optimization framework and analyze its convergence rate under commonly-used assumptions. We show a lower bound on the decrease i...

Let {Xn, n ≥ 1} be independent and identically distributed random variables with each Xn following skew normal distribution. Let Mn = max{Xk, 1 ≤ k ≤ n} denote the partial maximum of {Xn, n ≥ 1}. Liao et al. (2014) considered the convergence rate of the distribution of the maxima for random variables obeying the skew normal distribution under linear normalization. In this paper, we obtain the a...

We study two boosting algorithms, Coordinate Ascent Boosting and Approximate Coordinate Ascent Boosting, which are explicitly designed to produce maximum margins. To derive these algorithms, we introduce a smooth approximation of the margin that one can maximize in order to produce a maximum margin classifier. Our first algorithm is simply coordinate ascent on this function, involving a line se...

A distributed fixed-step power control algorithm is presented. It is a simple feedback adjustment algorithm using only local information. In the ideal case where there is no power constraint, it is guaranteed that existing users will not be dropped due to admission of new users. If it is infeasible to accommodate all of them, the new user will be blocked. When the constraint on maximum power is...

In this master thesis we explore skewed distributions extended from the normal distribution. A transition from the symmetric to an asymmetric setting is formulated for the univariate as well as the multivariate case. This is done by introducing an additional infinite-dimensional parameter, namely a skewing function π. Our main focus is on the extremal behavior of such skewed random vectors. For...

one of the problems associated with adaptive fir filters in the identification of systems with long impulse responses, is their excessive computational complexity. recently a new kind of adaptive filters, based on three-level clipping of the input signal has been presented for reduction of their computational complexity. in this paper, a theoretical analysis of the steady-state misalignment of ...