A Method Based on Interactive Evolutionary Computation for Increasing the Effectiveness of Advertisement Texts

Abstract

Interactive Evolutionary Computation (IEC) is used in this work in order to perform the optimization of several advertisement blocks of text. The advertisement texts follow a format similar to the one used in a technique called Article Spinning. This format allows an IEC algorithm to evolve the text, using words and phrases as variable parts which change according to the subjective evaluation of people interacting with the algorithm. After several generations, the IEC algorithm provides a version of the advertisement text that, in theory, should exhibit an increased performance, according to the subjective evaluation function it was evolved with. In order to demonstrate the efficiency of the texts, these are compared against a version determined by an expert in a field related to marketing. For this comparison, three tests are performed: recall, recognition, and persuasion tests. The results obtained show that IEC could effectively be used to increase the impact of an advertisement text.