Types of learning according to my personal understanding, the classification of learning methods in machine learning helps us face a specific problem, you can select an appropriate machine learning algorithm based on your goals. For example, to determine whether an email is a spam email, you need to use classification. To achieve the classification effect, you need to make the machine learn how to classify. This is the learning process. The learning process is divided into four categories: 1) Supervised Learning (supervised learning) 2) unsupervised learning (unsupervised learning) 3) Semi-Supervised Learning (semi-supervised learning) 4) reinforcement Learning)
The following is a conceptual distinction between the four types of learning: 1) Supervised Learning: each input sample in the training dataset, there is a label. For example, if a training dataset containing 1000 emails is specified, a corresponding y = {-} --- category problem occurs for each email, the number of classes divided into the specified classes --- regression problem, and a specific value is obtained.
2) unsupervised learning: the input training data is not marked with its category. Clustering is a common issue. If the data is news reports, you can divide the data into sports and financial categories based on the keywords in each report.
3) Semi-Supervised Learning: The training data is a mixture of labeled and unlabeled. Why not mark them all? In actual cases, the number of input data is too large. If all data is manually marked, the workload will be huge. The following figure shows the differences between the three learning methods.
4) enhanced learning: input data can stimulate the model to provide feedback. Feedback can be obtained not only from supervised learning, but also from rewards or punishments in the environment. For more information, see this article.
Learning methods in Machine Learning-types of learning
