Concept

Kernel (statistics)

The term kernel is used in statistical analysis to refer to a window function. The term "kernel" has several distinct meanings in different branches of statistics. In statistics, especially in Bayesian statistics, the kernel of a probability density function (pdf) or probability mass function (pmf) is the form of the pdf or pmf in which any factors that are not functions of any of the variables in the domain are omitted. Note that such factors may well be functions of the parameters of the pdf or pmf. These factors form part of the normalization factor of the probability distribution, and are unnecessary in many situations. For example, in pseudo-random number sampling, most sampling algorithms ignore the normalization factor. In addition, in Bayesian analysis of conjugate prior distributions, the normalization factors are generally ignored during the calculations, and only the kernel considered. At the end, the form of the kernel is examined, and if it matches a known distribution, the normalization factor can be reinstated. Otherwise, it may be unnecessary (for example, if the distribution only needs to be sampled from). For many distributions, the kernel can be written in closed form, but not the normalization constant. An example is the normal distribution. Its probability density function is and the associated kernel is Note that the factor in front of the exponential has been omitted, even though it contains the parameter , because it is not a function of the domain variable . The kernel of a reproducing kernel Hilbert space is used in the suite of techniques known as kernel methods to perform tasks such as statistical classification, regression analysis, and cluster analysis on data in an implicit space. This usage is particularly common in machine learning. In nonparametric statistics, a kernel is a weighting function used in non-parametric estimation techniques. Kernels are used in kernel density estimation to estimate random variables' density functions, or in kernel regression to estimate the conditional expectation of a random variable.

Official source

https://en.wikipedia.org/wiki/Kernel_(statistics)

About this result

This page is automatically generated and may contain information that is not correct, complete, up-to-date, or relevant to your search query. The same applies to every other page on this website. Please make sure to verify the information with EPFL's official sources.

Related courses (32)

MATH-342: Time series

A first course in statistical time series analysis and applications. ...

COM-406: Foundations of Data Science

We discuss a set of topics that are important for the understanding of modern data science but that are typically not taught in an introductory ML course. In particular we discuss fundamental ideas and techniques that come from probability, information the ...

COM-490: Large-scale data science for real-world data

This hands-on course covers tools and methods used by data scientists, from researching solutions to scaling prototypes on Spark clusters. Students engage with the full data engineering and data science pipeline, from data acquisition to extracting insight ...

Related lectures (32)

Feature Expansion and Kernel Methods

Explores feature expansion, kernel methods, SVM, and nonlinear classification in machine learning.

Data Representations: Learning Methods

Covers polynomial feature expansion, kernel functions, regression, and SVM, emphasizing the importance of choosing functions for feature expansion.

SVM Hyperparameters

Delves into SVM hyperparameters, showcasing the impact of C and kernel width on classification results.

Related publications (17)

Quantitative Convergence of Wasserstein Gradient Flows of Kernel Mean Discrepancies

We study the quantitative convergence of Wasserstein gradient flows of Kernel Mean Discrepancy (KMD) (also known as Maximum Mean Discrepancy (MMD)) functionals. Our setting covers in particular the training dynamics of shallow neural networks in the infinite-width and continuous time limit, as well as interacting particle systems with pairwise Riesz kernel interaction in the mean-field and overdamped limit.

arXiv2026

Kernel embeddings and the separation of measure phenomenon

We prove that kernel covariance embeddings lead to information-theoretically perfect separation of distinct continuous probability distributions. In statistical terms, we establish that testing for the equality of two nonatomic (Borel) probability measures on a locally compact uncountable Polish space is equivalent to testing for the singularity between two centered Gaussian measures on a reproducing kernel Hilbert space. The corresponding Gaussians are defined via the notion of kernel covariance embedding

National Academy of Sciences2026

Statistical and Algorithmic Foundations of Causal Learning and Decision-Making

Causal reasoning plays a central role in learning and decision-making in settings where distinguishing causal effect from spurious association is essential for valid prediction and action. However, many causal methods rely on strong assumptions, such as full observability, known causal structure, or independence across units. These assumptions are often violated in real-world applications. This thesis develops statistical and algorithmic foundations for causal learning and sequential decision-making, with

EPFL2026

Graph Chat

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.

Official source

https://en.wikipedia.org/wiki/Kernel_(statistics)

About this result

Related courses (32)

MATH-342: Time series

A first course in statistical time series analysis and applications. ...

COM-406: Foundations of Data Science

COM-490: Large-scale data science for real-world data

Related lectures (32)

Feature Expansion and Kernel Methods

Explores feature expansion, kernel methods, SVM, and nonlinear classification in machine learning.

Data Representations: Learning Methods

Covers polynomial feature expansion, kernel functions, regression, and SVM, emphasizing the importance of choosing functions for feature expansion.

SVM Hyperparameters

Delves into SVM hyperparameters, showcasing the impact of C and kernel width on classification results.

Related publications (17)

Quantitative Convergence of Wasserstein Gradient Flows of Kernel Mean Discrepancies

arXiv2026

Kernel embeddings and the separation of measure phenomenon

National Academy of Sciences2026

Statistical and Algorithmic Foundations of Causal Learning and Decision-Making

EPFL2026

Related people (30)

Rajita Ramesh Chandak

Expertise Nonparametric estimation, statistical inference, theory of machine learning Teaching & PhD Courses Nonparametric estimation and inference MATH-524 Nonparametric models are used to identify a wide range of relationships within data. This course gives a graduate-level overview of nonparametric statistical estimation and inference theory. Statistical inference MATH-562 Inference from the particular to the general based on probability models is central to the statistica

Devis Tuia

I come from Ticino and studied in Lausanne, between UNIL and EPFL. After my PhD at UNIL in remote sensing, I was postdoc in Valencia (Spain), Boulder (CO) and EPFL, working on model adaptation and prior knowledge integration in machine learning. In 2014 I became Research Assistant Professor at University of Zurich, where I started the 'multimodal remote sensing' group. In 2017, I joined Wageningen University (NL), where I was professor of the GeoInformation Science and Remote Sensing Laboratory. Since 2020, I joined EPFL Valais, to start the ECEO lab, working at the interface between Earth observation, machine learning and environmental sciences.

Emmanuel Abbé

Teaching & PhD PhD Students Denys Pushkin, Ortal Yona Senouf, Vladyslav Shashkov Past EPFL PhD Students Elisabetta Cornacchia, Aryo Lotfi Courses Reasoning in artificial intelligence MATH-700 Large language models have raised the potential of artificial intelligence in various applications, including science and mathematics, but their reasoning capabilities remain under investigations. This class focuses on defining, measuring and improving the reasoning capabilities of such AI models. Awa

Kamiar Aminian

Kamiar Aminian received the M.S. degree in electrical engineering in 1982, the Ph.D degree in biomedical engineering in 1989 and the Postgraduate degree on technical computing in 1993 from Ecole Polytechnique Fédérale de Lausanne (EPFL). He was assistant professor (1994-1996) with Sharif University of Technology (Tehran). He joint EPFL in 1996 where he is currently Professor of medical instrumentation and the director of the Laboratory of Movement Analysis and Measurement in the Institute of Bioengineering of EPFL. His research interests include methodologies for human movement monitoring and analysis in real world conditions mainly based on wearable technologies and inertial sensors with emphasis on gait, physical activity and sport. His research aims to perform outcome evaluation in orthopaedics, to improve motor function and intervention programs in aging and patients with movement disorders and pain, and to identify metrics of performance in sport science. Kamiar Aminian is a member of the International Society of Posture and Gait Research, the Institute of Electrical and Electronics Engineers, the European Society of Movement Analysis in Adults and Children, the Prevention of fall Network Europe, the Intentional Society of Biomechanics and the President of the 3D analysis of the human movement group. He is author or co-author of more than 450 scientific papers published in reviewed journals and presented at international conferences and holds 10 patents related to medical devices. Full CV ORCID

Yoshihito Kazashi

Samy Bengio

Expertise Large Language ModelsReasoningStatistical machine learningSpeech processingPerson authenticationMultimodal processingSequence processingArtificial neural networks, support vector machinesHidden markov models, gaussian mixture modelsetc Expertise Large Language ModelsReasoningStatistical machine learningSpeech processingPerson authenticationMultimodal processingSequence processingArtificial neural networks, support vector machinesHidden markov models, gaussian mixture modelsetc Mission Samy Bengio

Florent Gérard Krzakala

Martin Vetterli

Martin Vetterli was appointed president of EPFL by the Federal Council following a selection process conducted by the ETH Board, which unanimously nominated him. Professor Vetterli was born on 4 October 1957 in Solothurn and received his elementary and secondary education in Neuchâtel Canton. He earned a Bachelor’s degree in electrical engineering from ETH Zurich (ETHZ) in 1981, a Master’s of Science degree from Stanford University in 1982, and a PhD from EPFL in 1986. Professor Vetterli taught at Columbia University as an assistant and then associate professor. He was subsequently named full professor in the Department of Electrical Engineering and Computer Sciences at the University of California at Berkeley before returning to EPFL as a full professor at the age of 38. He has also taught at ETHZ and Stanford University. Professor Vetterli has earned numerous national and international awards for his research in electrical engineering, computer science and applied mathematics, including the National Latsis Prize in 1996. He is a fellow of both the Association for Computing Machinery and the Institute of Electrical and Electronics Engineers and a member the US National Academy of Engineering. He has published over 170 articles and three reference works. Professor Vetterli’s work on the theory of wavelets, which are used in signal processing, is considered to be of major importance by his peers, and his areas of expertise, including image and video compression and self-organized communication systems, are central to the development of new information technologies. As the founding director of the National Centre of Competence in Research on Mobile Information and Communication Systems, Professor Vetterli is a staunch advocate of transdisciplinary research. Professor Vetterli knows EPFL inside and out. An EPFL graduate himself, he began been teaching at the school in 1995, was vice president for International Affairs and then Institutional Affairs from 2004 to 2011, and served as dean of the School of Computer and Communication Sciences in 2011 and 2012. In addition to his role as president of the National Research Council of the Swiss National Science Foundation, a position he held from 2013 to 2016, he heads the EPFL’s Audiovisual Communications Laboratory (LCAV) since 1995. Professor Vetterli has supported more than 60 students in Switzerland and the United States in their doctoral work and makes a point of following their highly successful careers, whether it is in the academic or business world. He is the author of some 50 patents, some of which were the basis for start-ups coming out of his lab, such as Dartfish and Illusonic, while others were sold (e.g. Qualcomm) as successful examples of technology transfer. He actively encourages young researchers to market the results of their work.

Lenka Zdeborová

Lenka Zdeborová is a Professor of Physics and of Computer Science in École Polytechnique Fédérale de Lausanne where she leads the Statistical Physics of Computation Laboratory. She received a PhD in physics from University Paris-Sud and from Charles University in Prague in 2008. She spent two years in the Los Alamos National Laboratory as the Director's Postdoctoral Fellow. Between 2010 and 2020 she was a researcher at CNRS working in the Institute of Theoretical Physics in CEA Saclay, France. In 2014, she was awarded the CNRS bronze medal, in 2016 Philippe Meyer prize in theoretical physics and an ERC Starting Grant, in 2018 the Irène Joliot-Curie prize, in 2021 the Gibbs lectureship of AMS. She is an editorial board member for Journal of Physics A, Physical Review E, Physical Review X, SIMODS, Machine Learning: Science and Technology, and Information and Inference. Lenka's expertise is in applications of concepts from statistical physics, such as advanced mean field methods, replica method and related message-passing algorithms, to problems in machine learning, signal processing, inference and optimization. She enjoys erasing the boundaries between theoretical physics, mathematics and computer science.

Michaël Unser

Expertise Image Processing, Medical Imaging, Biological Imaging, Wavelets, Splines, Multiresolution Michael Unser is Full Professor with EPFL’s school of Engineering and the academic director of EPFL's Center for Imaging, Lausanne, Switzerland. His primary areas of investigation are signal processing, biomedical imaging and applied functional analysis. He is internationally recognized for his research contributions to sampling theory, wavelets, the use of splines for image processing, stochastic processes,

Related units (13)

Biomedical Imaging Laboratory

CIBM - Signal Processing

IMAGING - Administration

Related concepts (9)

Kernel smoother

A kernel smoother is a statistical technique to estimate a real valued function as the weighted average of neighboring observed data. The weight is defined by the kernel, such that closer points are given higher weights. The estimated function is smooth, and the level of smoothness is set by a single parameter. Kernel smoothing is a type of weighted moving average. Let be a kernel defined by where: is the Euclidean norm is a parameter (kernel radius) D(t) is typically a positive real valued function, whose value is decreasing (or not increasing) for the increasing distance between the X and X0.

Kernel density estimation

In statistics, kernel density estimation (KDE) is the application of kernel smoothing for probability density estimation, i.e., a non-parametric method to estimate the probability density function of a random variable based on kernels as weights. KDE answers a fundamental data smoothing problem where inferences about the population are made, based on a finite data sample. In some fields such as signal processing and econometrics it is also termed the Parzen–Rosenblatt window method, after Emanuel Parzen and Murray Rosenblatt, who are usually credited with independently creating it in its current form.

Kernel method

In machine learning, kernel machines are a class of algorithms for pattern analysis, whose best known member is the support-vector machine (SVM). These methods involve using linear classifiers to solve nonlinear problems. The general task of pattern analysis is to find and study general types of relations (for example clusters, rankings, principal components, correlations, classifications) in datasets.