A Generalized View on Continuum Regression

ABSTRACT. We generalize the relationship between continuum regression (Stone & Brooks, 1990) and ridge regression, by showing that any optimization principle will yield a regressor proportional to a ridge regressor, provided only that the principle implies maximizing a function of the regressor's sample correlation coefficient and its sample variance. This relationship shows that continuum regression as defined via ridge regression (least squares ridge regression) is a more generally valid methodology than previously realized, and also opens up for alternative choices of its second and subsequent factors.     

Combinatorics of TCP reordering

We study a combinatorial problem motivated by a receiver-oriented model of TCP traffic from Istrate et al. (2006), that incorporates information on both arrival times, and the dynamics of packet IDs. An important component of this model is a many-to-one mapping FB from sequences of IDs into a sequence of buffer sizes. We show that: i) Given a buffer sequence B, constructing a sequence A of IDs that belongs to the preimage of B is no harder than finding matchings in bipartite graph. ii) Counting the number of sequences A of packet IDs that belong to the preimage of B can be done in linear time in the special case when there exists a constant upper bound on the maximum entry in B. iii) This problem also has a fully polynomial randomized approximation scheme when we have a constant upper bound on the number of repeats in the packet sequences in the preimage. We also provide experimental evidence that the two previous results suffice to efficiently count the number of preimages for buffer sequences observed in real TCP data.     

Wage Setting in Sweden: An Empirical Test of a Barebones Union Model

ABSTRACT In this paper a simple monopoly union model for the Swedish economy is presented and estimated. Interest is focused on the union's objectives. The results indicate that two popular representations of union objectives — wage bill and rent maximisation — are rejected by the data. Instead, preferences are highly concave in the wage rate, with an estimated degree of relative risk aversion well above unity evaluated at sample means.