Skip to Content

Machine Learning

There are numerous applications for machine learning including adaptive websites, marketing, and search engines. Machine learning algorithms can be trained using sample data in order to make decisions or predictions without being specifically programmed for the task.

Approaches to machine learning vary. Supervised learning and semi-supervised learning algorithms use training data with both inputs and desired outputs. This type of algorithm is typically used in classification systems and recommender systems. Unsupervised learning algorithms only receive inputs as training data and look for structure within. Reinforcement learning algorithms are used when exact mathematical models are not feasible.

The performance of these algorithms can be improved through the application of various processes. Feature learning algorithms preprocess data and transform it in order to be more useful in classifying or performing predictions. Sparse dictionary learning is a feature learning method used for processing various signals including audio and video. 

Anomaly detection finds exceptions or outliers in a set of data. It can be used in diverse fields such as uncovering bank fraud or finding text errors. Decision tree learning uses a model that predicts a target variable’s value based on several input variables. Association rule learning is a method for discovering relationships between variables in large sets of data such as in natural language processing.

There are four major models of machine learning. Artificial neural networks are modeled after the biological networks of animal brains and are used for deep learning. Support vector machines consist of related supervised learning methods used for the classification of data. Bayesian networks are used in areas as diverse as modeling speech signals and representing the relationship between symptoms and disease. Genetic algorithms are a technique that imitates the process of natural selection.

7 Steps of Machine Learning