Machine Learning

Machine learning is the scientific area where concepts and techniques from statistics, mathematics, and computer science are used in order to find knowledge and insights from data using computers.

With the rapid increase of available data and computer power, machine learning techniques have during the last decade become more and more popular tools when analyzing large amounts of data. The traditional way of performing analysis on data is to define some appropriate model that experts believe represents the underlying description of the data and then try to validate that belief. In machine learning an important difference compared to the traditional way is that one lets the machine (computer) itself find the appropriate model that describes the data best.

Examples of important areas within machine learning are pattern recognition, clustering, classification, anomaly detection, and reinforcement learning. Machine learning also contains interesting procedures for performing regression analysis, which extend the ideas of ordinary least squares regression by introducing nonlinearities or regularizations.

FCC has worked with several partners in this field, developing machine learning algorithms and procedures; examples include analysis of statistical properties of processes to ensure desired quality, development of tools to be used for optimal product development, and working with partners delivering a specific measurement instrument in order to develop analysis methods and products using that instrument.

© 2014-2017  Fraunhofer-Chalmers Centre