As method of machine learning, supervised learning makes computers learn data, together with the right answers prepared by humans, so that they become able to classify data in a way intended by humans. The learning process uses "training data," or teaching sets of data with labels for classification attached by humans.
The method, which is effective in cases where the information for computers to output is already determined, is expected to be used in such areas as detection of wrongdoing, diagnosis of diseases and weather forecasting, in which forecasts are made about what will happen in the future based on past cases. Moreover, as humans teach right answers, the method features high accuracy and quickness of learning. Disadvantages of supervised learning includes that it is not applicable to areas where no right answer exists, and that success in learning depends on the quality of the training data.
