Solving a bi-criteria integer linear programming model to enhance tweet contextualization
Fernando Paredes, Santiago de Chile, Javier Pereira · 2012
We propose a bi-criteria integer linear programming approach to support constructive searching of non-dominated document passage assemblies in tweet contextualization, formulating this problem in terms of coverage of a given set of interesting terms in a tweet taken from Twitter, with the minimum number of passages and minimum redundancy in a feasible assembly. We show that resolution of this problem could be very efficient, without the disadvantages of problem size or sparse information data structure, as found in precedent works, where the same formulation was proposed.