Modèle statistiqueUn modèle statistique est une description mathématique approximative du mécanisme qui a généré les observations, que l'on suppose être un processus stochastique et non un processus déterministe. Il s’exprime généralement à l’aide d’une famille de distributions (ensemble de distributions) et d’hypothèses sur les variables aléatoires X1, . . ., Xn. Chaque membre de la famille est une approximation possible de F : l’inférence consiste donc à déterminer le membre qui s’accorde le mieux avec les données.
Similarity measureIn statistics and related fields, a similarity measure or similarity function or similarity metric is a real-valued function that quantifies the similarity between two objects. Although no single definition of a similarity exists, usually such measures are in some sense the inverse of distance metrics: they take on large values for similar objects and either zero or a negative value for very dissimilar objects. Though, in more broad terms, a similarity function may also satisfy metric axioms.
Loi multinomialeEn théorie des probabilités, la loi multinomiale (aussi appelée distribution polynomiale) généralise la loi binomiale. Tandis que la loi binomiale concerne le nombre de succès lors d'une série de n épreuves de Bernoulli indépendantes, comme dans le jeu de pile ou face, la loi multinomiale ne se restreint pas aux épreuves comportant deux issues. La loi multinomiale s'applique par exemple au cas de n jets d'un dé à six faces : l'apparition du seul peut être modélisé par une loi binomiale alors que l'ensemble des apparitions des à 6 est modélisé par une loi multinomiale.
Fisher information metricIn information geometry, the Fisher information metric is a particular Riemannian metric which can be defined on a smooth statistical manifold, i.e., a smooth manifold whose points are probability measures defined on a common probability space. It can be used to calculate the informational difference between measurements. The metric is interesting in several respects. By Chentsov’s theorem, the Fisher information metric on statistical models is the only Riemannian metric (up to rescaling) that is invariant under sufficient statistics.
Additive smoothingIn statistics, additive smoothing, also called Laplace smoothing or Lidstone smoothing, is a technique used to smooth categorical data. Given a set of observation counts from a -dimensional multinomial distribution with trials, a "smoothed" version of the counts gives the estimator: where the smoothed count and the "pseudocount" α > 0 is a smoothing parameter. α = 0 corresponds to no smoothing. (This parameter is explained in below.
TF-IDFLe TF-IDF (de l'anglais term frequency-inverse document frequency) est une méthode de pondération souvent utilisée en recherche d'information et en particulier dans la fouille de textes. Cette mesure statistique permet d'évaluer l'importance d'un terme contenu dans un document, relativement à une collection ou un corpus. Le poids augmente proportionnellement au nombre d'occurrences du mot dans le document. Il varie également en fonction de la fréquence du mot dans le corpus.
Semantic lexiconA semantic lexicon is a digital dictionary of words labeled with semantic classes so associations can be drawn between words that have not previously been encountered. Semantic lexicons are built upon semantic networks, which represent the semantic relations between words. The difference between a semantic lexicon and a semantic network is that a semantic lexicon has definitions for each word, or a "gloss". Semantic lexicons are made up of lexical entries. These entries are not orthographic, but semantic, eliminating issues of homonymy and polysemy.
Cache language modelA cache language model is a type of statistical language model. These occur in the natural language processing subfield of computer science and assign probabilities to given sequences of words by means of a probability distribution. Statistical language models are key components of speech recognition systems and of many machine translation systems: they tell such systems which possible output word sequences are probable and which are improbable.
Scoring algorithmScoring algorithm, also known as Fisher's scoring, is a form of Newton's method used in statistics to solve maximum likelihood equations numerically, named after Ronald Fisher. Let be random variables, independent and identically distributed with twice differentiable p.d.f. , and we wish to calculate the maximum likelihood estimator (M.L.E.) of . First, suppose we have a starting point for our algorithm , and consider a Taylor expansion of the score function, , about : where is the observed information matrix at .
