Ciência de dados aplicada à aquisição da língua oral

Abstract

The goal of this work is to help in the understanding of the acquisition of oral verbal language process, from the exploration of a speech database, using principles, techniques and tools of Data Science. A linguistic database, composed of results of oral production tests applied to children from 3 to 7 years old, was prepared and subjected to Data Warehouse techniques, Business Intelligence tools and data mining. K-Means, Apriori and Naïve Bayes were applied. Among the results obtained, it is worth mentioning that: the marked sociolinguistic variation of the researched region is being less performed by the new generations; around the age of 5, there is a transition phase in which important discrepancies occur in the child's acquisition of the phonological system; at the beginning of literacy, there are cases of children who do not pronounce all the sounds of Portuguese; the first steps were taken towards a model that relates determinant characteristics of honoarticulatory maturation. It concludes that, despite some difficulties, it is possible to process linguistic data satisfactorily in data warehouse modeling, in BI tools and in data mining algorithms.