Text Mining and Statistical Learning for the Analysis of the Voice of the Customer

This paper analyzes the content of texts through a Text Mining classification model for the particular case of the Tweets made about the Miniso brand in Mexico during the period from November 17 to 24, 2018. The analysis involves the extraction of the data, the cleaning of the text and supervised support models for high-dimensional data, obtaining as a result the classification of the tweets in the topics: Positive, Negative, Advertising or Requirements of new Branches. As well as the use of resampling techniques to measure the variability of the performance of the model and to improve the accuracy of the parameters. This practice allows to reduce time spent reading texts, especially in Social Networks, finding faster and more efficient trends that help decision-making and respond quickly to customer demand. © 2020, Springer Nature Switzerland AG.