Presidential elections in Peru: text mining of La República newspaper publishers

Abstract

It analyzes the editorials of the Peruvian newspaper La República published from May 5 to June 6, 2021, dates in which the political campaigns for the second round of elections for the presidency of Peru took place. The aim is to discover relevant information to understand the use of the terminologies communicated through the written language in that newspaper. Using the text mining technique and machine learning algorithms for the analysis of unstructured data, a dendrogram and clusters were created to validate and estimate the groupings of the most frequent terms. In relation to each candidate, the regression analysis shows a differentiated association in the use of words. These words form a network that expresses the political conjuncture of the period under study and is reinforced by the construction of a cloud of words with the highest frequency of use. It is concluded that the media construct their own social representations on various local and national issues in a way that consciously or unconsciously evidences their preferences. These preferences are not foreign in the editorials of the newspaper La República. The importance of conducting this type of study lies in the potential of the theory of social representations to place the researcher in front of the position held by the subjects, who construct and communicate the antagonistic onslaughts of the national social reality.