This work aims to analyze the convergence of the K-means method, an unsupervised learning
algorithm that groups n data into k-clusters. In this sense, we presented some of the advantages
and disadvantages of the K-means method, comparing the original clustering and the
clustering done by the algorithm. Also, we presented the application of the algorithm in two
data sets: breast cancer and diabetes, analyzing the clustering done by K-means as well as the
patterns and regularities present in the clusters. In this way, we seek to present an introductory
study of Machine Learning theory, which seeks to make machines perform tasks without
being instructed all the time, starting only from some initial instructions. Specifically, we seek
to understand some of its definitions and characteristics that will allow identifying the type of
learning studied.
|
