Optimization Algorithm for Improving the Efficacy of an Information Retrieval Model

The aim of Information Retrieval (IR) is to nd the documents that are more relevant for a query, usually given by a user. This problem is very di cult, and in the last four decades a lot of di erent models were proposed, the most famous being the logical models, the vector space models, and the probabilistic models. In this paper is proposed a greedy algorithm for maximizing the e cacy of an Information Retrieval model based on Discrete Fourier Transform (DFT), which has shown a good e cacy level in the rst tests. Even if the mathematical programming model used to increase the e cacy is a Mixed-Integer Nonlinear Program (MINLP), with nonlinear objective function and binary variables, its structure is very simple and a greedy algorithm can nd the optimal solution.