Paul H. Kvam, Brani Vidaković, Seong-joon Kim · Wiley series in probability and statistics · 2022
This chapter presents a brief exposition of classification and statistical learning that can be used in machine learning, discriminant analysis, pattern recognition, neural networks, and data mining. Nonparametric methods play a vital role in statistical learning. Discriminant analysis is the statistical name for categorical prediction. In terms of bias versus variance, a linear classification model represents a strict global model with potential for bias, but low variance that makes the classifier more stable. The nearest-neighbor methods can be used to create nonparametric regressions by determining the regression curve. The perceptron is overly simplistic and has numerous shortcomings, but it also represents the first neural network. A key in training the network is to find the weights to go along with the activation functions that lead to supervised learning. Binary trees offer a graphical and logical basis for empirical classification.
Cited by 0
Claude Sammut, Geoffrey I. Webb · Encyclopedia of Machine Learning · 2010
Cited by 0
Thiessen, Erik · Figshare · 2018
Department of Psychology
Cited by 0Open access
Victor M. Zavala · Cambridge University Press eBooks · 2025
Build a firm foundation for studying statistical modelling, data science, and machine learning with this practical introduction to statistics, written with chemical engineers in mind. It introduces a data–model–decision approach to applying statistical methods to real-world chemical engineering challenges, establishes links between statistics, probability, linear algebra, calculus, and optimization, and covers classical and modern topics such as uncertainty quantification, risk modelling, and decision-making under uncertainty. Over 100 worked examples using Matlab and Python demonstrate how to apply theory to practice, with over 70 end-of-chapter problems to reinforce student learning, and key topics are introduced using a modular structure, which supports learning at a range of paces and levels. Requiring only a basic understanding of calculus and linear algebra, this textbook is the ideal introduction for undergraduate students in chemical engineering, and a valuable preparatory text for advanced courses in data science and machine learning with chemical engineering applications.
Cited by 0
Wolfgang Garn · 2024
Cited by 0
William FordeThompson · Music in the Social and Behavioral Sciences: An Encyclopedia · 2014
Cited by 0
Claude Sammut, Geoffrey I. Webb · Encyclopedia of Machine Learning and Data Mining · 2017
Cited by 0
Robin Lovelace, Jakub Nowosad, Jannes Muenchow · 2025
Cited by 0
Dirk P. Kroese, Zdravko I. Botev, Thomas Taimre et al. · 2019
Cited by 0
Anna C. Schapiro, Nicholas B. Turk‐Browne · Elsevier eBooks · 2015
Cited by 101
Rebuschat, Patrick · Lancaster EPrints (Lancaster University) · 2012
Cited by 0
Yiming Yang, Jaime G. Carbonell, Ralf D. Brown et al. · 2014
Translingual information retrieval (TLIR) consists of providing a query in one language and searching document collections in one or more dierent languages. This paper introduces new TLIR methods and reports on comparative TLIR experiments with these new methods and with previously reported ones in a realistic setting. Methods fall into two categories: query translation and statistical-IR approaches establishing translingual associations. The results show that using bilingual corpora for automated extraction of term equivalences in context outperforms dictionary-based methods. Translingual versions of the Generalized Vector Space Model (GVSM) and Latent Semantic Indexing (LSI) also perform well, as does translingual pseudo relevance feedback (PRF) and Example-Based Term-in-context Translation (EBT). All showed relatively small performance loss between monolingual and translingual versions, rang-ing between 87 % to 101 % of monolingual IR performance. Query translation based on a general machine-readable bilingual dictionary { heretofore the most popular method { did not match the performance of other, more sophisticated methods. Also, the previous very high LSI results in the literature based on \\mate-nding"were superseded by more realistic relevance-based eval-uations; LSI performance proved comparable to that of other statistical corpus-based methods. Keywords: Information retrieval
Cited by 0
Thiessen, Erik · KiltHub Repository · 2018
Department of Psychology
Cited by 0Open access
2014
In this chapter, we will describe some techniques to learn from data, and to make a prediction based on a set of features. We will use a training set, where those features were observed, aswhere the variable of interest might be• Whether an insured will buy additional (optional) coverage, or not • Whether a claimant will be represented by an attorney, or not (see e.g. the automobileinjury insurance claims in Frees (2009))• Whether an insured will have some specific disease, or not • Whether a loaner will be considered a good or a bad client (in this chapter)All the techniques mentioned in this chapter will be used on a binary variable of interest (good or bad client), but one can easily extend most of them to an ordered discrete variable of interest.
Cited by 0
WORLD SCIENTIFIC eBooks · 2025
Cited by 0
WORLD SCIENTIFIC eBooks · 2024
Cited by 0
Erik D. Thiessen · Cambridge University Press eBooks · 2012
Cited by 5
Vladimir Vapnik, Rauf Izmailov · Machine Learning · 2018
Cited by 66Open access
Peyrard, Maxime, Cho, Kyunghyun · SPIRE - Sciences Po Institutional REpository · 2025
<div> This work demonstrates that the tools and principles driving the success of large language models (LLMs) can be repurposed to tackle distributionlevel tasks, where the goal is to predict properties of the data-generating distribution rather than labels for individual datapoints. These tasks encompass statistical inference problems such as parameter estimation, hypothesis testing, or mutual information estimation. Framing these tasks within traditional machine learning pipelines is challenging, as supervision is typically tied to individual datapoint. We propose meta-statistical learning, a framework inspired by multi-instance learning that reformulates statistical inference tasks as supervised learning problems. In this approach, entire datasets are treated as single inputs to neural networks, which predict distributionlevel parameters. Transformer-based architectures, without positional encoding, provide a natural fit due to their permutation-invariance properties. By training on large-scale synthetic datasets, meta-statistical models can leverage the scalability and optimization infrastructure of Transformer-based LLMs. We demonstrate the framework's versatility with applications in hypothesis testing and mutual information estimation, showing strong performance, particularly for small datasets where traditional neural methods struggle. </div>
Cited by 0Open access
Jenny R. Saffran · Current Directions in Psychological Science · 2003
What types of mechanisms underlie the acquisition of human language? Recent evidence suggests that learners, including infants, can use statistical properties of linguistic input to discover structure, including sound patterns, words, and the beginnings of grammar. These abilities appear to be both powerful and constrained, such that some statistical patterns are more readily detected and used than others. Implications for the structure of human languages are discussed.
Cited by 623