Use of Artificial Intelligence in Brazil Mortality Data Analysis

The exponential growth of health data in the world is evident. Thus, traditional analysis of variables is not viable due to the large volume. The use of artificial intelligence in the exploratory analysis of mortality data from Brazil through unsupervised machine learning algorithms is proposed. The basis used for the study was a modified DATASUS mortality database, in the interval between 2007 and 2017. After pre-processing, of the 13.040.656 lines and 166 variables from the original base, 3.021.787 lines and 36 variables remained. Using Konohen’s self-organizing map algorithm, component maps of the final variables were extracted. The insights extracted directly from the comparison between the maps were confirmed by a statistical analysis of the data. Therefore, it is notable that the use of such algorithms in the exploratory analysis of data, even in Big Data scenarios, become extremely recommended for the creation of guiding data studies.