| Popis: |
The aim of this paper is to present advanced methods for the search for new knowledge contained in BIG DATA, huge, growing datasets and technology. To the end, machine learning is illuminated, a science that trains computers to analyze data and solve tasks without having to be explicitly programmed to do it. Sub-areas of machine learning, as part of artificial inteligence, are presented to solve these problems. The areas of classical machine learning are suppervised learning (classification and regression), unsupervised learning (clistering, pattern search, dimensionality reduction), the support machine vector, and decision tree. The areas of modern machine learning are enhanced learning, ensemble methods, neural networks and deep learning, and Bayesian networks, as special additional sub-fields and methods in the field of machine learning. The work and results of this paper are significant because the decribed machine learning methods are inevitable, with a tendency to absorb new datafrom many sources, creating new sources of knowledge. Web 2.0, with Google apps, blogs, wikipedia, social networks, Facebook, folksonomies, videosharing online, Web mobile apps, is just one of these inexhaustible sources of data. Knowledge is obtained from data analysis. |
