DeepL is a German AI research company known for its language AI platform, which includes DeepL Translator and DeepL Voice, and for DeepL Agent, an AI agent capable of planning workflows and using office systems and tools autonomously, in response to natural language instructions. Its algorithm uses the transformer architecture. It offers a paid subscription for additional features and access to its translation application programming interface. DeepL was founded in 2017 by Jaroslaw Kutylowski and is a unicorn, valued at $2 billion after a Series C funding round raised $300 million in May 2024. Its more than 200,000 business customers include a large proportion of the Fortune 500. == History == The translating system was first developed within Linguee by a team led by Chief Technology Officer Jarosław Kutyłowski in 2016. It was launched as DeepL Translator on 28 August 2017 and offered translations between English, German, French, Spanish, Italian, Polish and Dutch. At its launch, it claimed to have surpassed its competitors in blind tests and BLEU scores, including Google Translate, Amazon Translate, Microsoft Translator and Facebook's translation feature. With the release of DeepL in 2017, Linguee's company name was changed to DeepL GmbH, and it is also financed by advertising on its sister site, linguee.com. Support for Portuguese and Russian was added on 5 December 2018. In July 2019, Jarosław Kutyłowski became the CEO of DeepL GmbH and restructured the company into a Societas Europaea in 2021. Translation software for Microsoft Windows and macOS was released in September 2019. Support for Chinese (simplified) and Japanese was added on 19 March 2020, which the company claimed to have surpassed the aforementioned competitors as well as Baidu and Youdao. Then, 13 more European languages were added in March 2021: Bulgarian, Czech, Danish, Estonian, Finnish, Greek, Hungarian, Latvian, Lithuanian, Romanian, Slovak, Slovenian, and Swedish, bringing the total number of supported languages to 24. On 25 May 2022, support for Indonesian and Turkish was added, and support for Ukrainian was added on 14 September 2022. In January 2023, the company reached a valuation of 1 billion euro and became the most valued startup company in Cologne. At the end of the month, support for Korean and Norwegian (Bokmål) was also added. In May 2024, the company announced an investment of US$300 million at AI. In January 2026, more languages were supported, including Luxembourgish and Irish. == Services == === Translation method === The service uses a proprietary algorithm with convolutional neural networks (CNNs) that have been trained with the Linguee database. According to the developers, the service uses a newer improved architecture of neural networks, resulting in a more natural sound of translations than by competing services. The translation is generated using a supercomputer that reaches 5.1 petaflops and is operated in Iceland with hydropower. DeepL's data centers are located at the EcoDataCenter in Falun, Sweden, which is a data center for sustainability. In general, CNNs are slightly more suitable for long coherent word sequences, but they have so far not been used by the competition because of their weaknesses compared to recurrent neural networks. The weaknesses of DeepL are compensated for by supplemental techniques, some of which are publicly known. === Translator and subscription === The translator can be used for free with a maximum limit of 1,500 characters per translation. Microsoft Word and PowerPoint files in Office Open XML file formats (.docx and .pptx) and PDF files up to 5MB in size can also be translated. It offers paid subscription DeepL Pro, which has been available since March 2018 and includes application programming interface access and a software plug-in for computer-assisted translation tools, including SDL Trados Studio. Unlike the free version, translated texts are stated to not be saved on the server; also, the character limit is removed. The monthly pricing model includes a set amount of text, with texts beyond that being calculated according to the number of characters. ==== Supported languages ==== As of May 2026, the translation service supports the following languages: Additionally, these languages are currently in beta, indicated by an asterisk after their name in the language picker: === DeepL Write === In November 2022, DeepL launched a tool to improve monolingual texts in English and German, called DeepL Write. In December, the company removed access and informed journalists that it was only for internal use and that DeepL Write would be relaunched in early 2023. The public beta version was then released on January 17, 2023. In the summer of 2024, DeepL announced the availability of two more languages in DeepL Write: French and Spanish. By January 2024, DeepL had added an additional two: Portuguese (European and Brazilian) and Italian. === DeepL Agent === In November 2025, DeepL launched an AI agent called DeepL Agent which is capable of operating business applications in a human-like manner. == Reception == The reception of DeepL has been generally positive. TechCrunch appreciates it for the accuracy of its translations and stating that it was more accurate and nuanced than Google Translate. Le Monde thanks its developers for translating French text into more "French-sounding" expressions. RTL Z stated that DeepL Translator "offers better translations […] when it comes to Dutch to English and vice versa". La Repubblica, and a Latin American website, "WWWhat's new?", showed praise as well. A 2018 paper by the University of Bologna evaluated the Italian-to-German translation capabilities and found the preliminary results to be similar in quality to Google Translate. In September 2021, Slator remarked that the language industry response was more measured than the press and noted that DeepL is still highly regarded by users. A reviewer noted in 2018 that DeepL had far fewer languages available for translation than competing products. == Awards and honors == DeepL won the 2020 Webby Award for Best Practices and the 2020 Webby Award for Technical Achievement (Apps, Mobile, and Features), both in the category Apps, Mobile & Voice. In April 2025, DeepL was featured in the Forbes AI 50 list.
SWIG
The Simplified Wrapper and Interface Generator (SWIG) is an open-source software tool used to connect computer programs or libraries written in C or C++ with scripting languages such as Lua, Perl, PHP, Python, R, Ruby, Tcl, and other language implementations like C#, Java, JavaScript, Go, D, OCaml, Octave, Scilab and Scheme. Output can also be in the form of XML. == Function == The aim is to allow the calling of native functions (that were written in C or C++) by other programming languages, passing complex data types to those functions, keeping memory from being inappropriately freed, inheriting object classes across languages, etc. The programmer writes an interface file containing a list of C/C++ functions to be made visible to an interpreter. SWIG will compile the interface file and generate code in regular C/C++ and the target programming language. SWIG will generate conversion code for functions with simple arguments; conversion code for complex types of arguments must be written by the programmer. The SWIG tool creates source code that provides the glue between C/C++ and the target language. Depending on the language, this glue comes in three forms: a shared library that an extant interpreter can link to as some form of extension module, or a shared library that can be linked to other programs compiled in the target language (for example, using Java Native Interface (JNI) in Java). a shared dynamic library source code that should be compiled and dynamically loaded (e.g. Node.js native extensions) SWIG is not used for calling interpreted functions by native code; this must be done by the programmer manually. == Example == SWIG wraps simple C declarations by creating an interface that closely matches the way in which the declarations would be used in a C program. For example, consider the following interface file: In this file, there are two functions sin() and strcmp(), a global variable Foo, and two constants STATUS and VERSION. When SWIG creates an extension module, these declarations are accessible as scripting language functions, variables, and constants respectively. In Python: == Purpose == There are two main reasons to embed a scripting engine in an existing C/C++ program: The program can then be customized far faster, via a scripting language instead of C/C++. The scripting engine may even be exposed to the end-user, so that they can automate common tasks by writing scripts. Even if the final product is not to contain the scripting engine, it may nevertheless be very useful for writing test scripts. There are several reasons to create dynamic libraries that can be loaded into extant interpreters, including: Provide access to a C/C++ library which has no equivalent in the scripting language. Write the whole program in the scripting language first, and after profiling, rewrite performance-critical code in C or C++. == History == SWIG is written in C and C++ and has been publicly available since February 1996. The initial author and main developer was David M. Beazley who developed SWIG while working as a graduate student at Los Alamos National Laboratory and the University of Utah and while on the faculty at the University of Chicago. Development is currently supported by an active group of volunteers led by William Fulton. SWIG has been released under a GNU General Public License. == Google Summer of Code == SWIG was a successful participant of Google Summer of Code in 2008, 2009, 2012. In 2008, SWIG got four slots. Haoyu Bai spent his summers on SWIG's Python 3.0 Backend, Jan Jezabek worked on Support for generating COM wrappers, Cheryl Foil spent her time on Comment 'Translator' for SWIG, and Maciej Drwal worked on a C backend. In 2009, SWIG again participated in Google Summer of Code. This time four students participated. Baozeng Ding worked on a Scilab module. Matevz Jekovec spent time on C++0x features. Ashish Sharma spent his summer on an Objective-C module, Miklos Vajna spent his time on PHP directors. In 2012, SWIG participated in Google Summer of Code. This time four out of five students successfully completed the project. Leif Middelschulte worked on a C target language module. Swati Sharma enhanced the Objective-C module. Neha Narang added the new module on JavaScript. Dmitry Kabak worked on source code documentation and Doxygen comments. == Alternatives == For Python, similar functionality is offered by SIP, Pybind11, and Boost's Boost.python library. == Projects using SWIG == ZXID (Apache License, Version 2.0) Symlabs SFIS (commercial) LLDB GNU Radio up to (including) version 3.8.x.x; later versions use Pybind11 Xapian TensorFlow Apache SINGA QuantLib Babeltrace
Persian Speech Corpus
The Persian Speech Corpus is a Modern Persian speech corpus for speech synthesis. The corpus contains phonetic and orthographic transcriptions of about 2.5 hours of Persian speech aligned with recorded speech on the phoneme level, including annotations of word boundaries. Previous spoken corpora of Persian include FARSDAT, which consists of read aloud speech from newspaper texts from 100 Persian speakers and the Telephone FARsi Spoken language DATabase (TFARSDAT) which comprises seven hours of read and spontaneous speech produced by 60 native speakers of Persian from ten regions of Iran. The Persian Speech Corpus was built using the same methodologies laid out in the doctoral project on Modern Standard Arabic of Nawar Halabi at the University of Southampton. The work was funded by MicroLinkPC, who own an exclusive license to commercialise the corpus, though the corpus is available for non-commercial use through the corpus' website. It is distributed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. The corpus was built for speech synthesis purposes, but has been used for building HMM based voices in Persian. It can also be used to automatically align other speech corpora with their phonetic transcript and could be used as part of a larger corpus for training speech recognition systems. == Contents == The corpus is downloadable from its website, and contains the following: 396 .wav files containing spoken utterances 396 .lab files containing text utterances 396 .TextGrid files containing the phoneme labels with time stamps of the boundaries where these occur in the .wav files. phonetic-transcript.txt which has the form "[wav_filename]" "[Phoneme Sequence]" in every line orthographic-transcript.txt which has the form "[wav_filename]" "[Orthographic Transcript]" in every line
Receptron
The receptron (short for "reservoir perceptron") is a neuromorphic data processing model — specifically neuromorphic computing — that generalizes the traditional perceptron, by incorporating non-linear interactions between inputs. Unlike classical perceptron, which rely on linearly independent weights, the receptron leverages complexity in physical substrates, such as the electric conduction properties of nanostructured materials or optical speckle fields, to perform classification tasks. The receptron bridges unconventional computing and neural network principles, enabling solutions that do not require the training approaches typical of artificial neural networks based on the perceptron model. == Algorithm == The receptron is an algorithm for supervised learning of binary classifiers, so a classification algorithm that makes its predictions based on a predictor function, combining a set of weights with the feature vector. The mathematical model is based on the sum of inputs with non-linear interactions: S = ∑ k = 1 n x j w ~ j ( x → ) | S ∈ R {\displaystyle S=\sum _{k=1}^{n}x_{j}{\widetilde {w}}_{j}({\vec {x}})|S\in R} (1) where j ∈ [ 1 , n ] {\displaystyle j\in [1,n]} and w ~ j {\displaystyle {\widetilde {w}}_{j}} are non-linear weight functions depending on the inputs, x → {\displaystyle {\vec {x}}} . Nonlinearity will typically make the system extremely complex, and allowing for the solution of problems not solvable through the simpler rules of a linear system, such as the perceptron or McCulloch Pitts neurons, which is based on the sum of linearly independent weights: S = ∑ k = 1 n x j w j p {\displaystyle S=\sum _{k=1}^{n}x_{j}w_{j}^{p}} (2) where w j {\displaystyle w_{j}} are constant real values. A consequence of this simplicity is the limitation to linearly separable functions, which necessitates multi-layer architectures and training algorithms like backpropagation As in the perceptron case, the summation in Eq. 1 origins the activation of the receptron output through the thresholding process, Y ( x 1 , . . . , x n ) = { 1 if S > th 0 if S ≤ th {\displaystyle Y(x_{1},...,x_{n})={\begin{cases}1&{\text{if }}S>{\text{th}}\\0&{\text{if }}S\leq {\text{th}}\end{cases}}} (3) where th is a constant threshold parameter. Equation 3 can be written by using the Heaviside step function. The weight functions w ~ ( x → ) {\displaystyle {\widetilde {w}}({\vec {x}})} can be written with a finite number of parameters w j 1 . . . j n {\displaystyle w_{j_{1}...j_{n}}} , simplifying the model representation. One can Taylor-expand w ~ ( x → ) {\displaystyle {\widetilde {w}}({\vec {x}})} and use the idempotency of Boolean variables ( x j ) q = x j ∀ q ≥ 1 {\displaystyle (x_{j})^{q}=x_{j}\forall q\geq 1} such that S ′ = b + ∑ k = 1 n x j w ~ j ( x → ) {\displaystyle S'=b+\sum _{k=1}^{n}x_{j}{\widetilde {w}}_{j}({\vec {x}})} can be written as S ′ ( x → ) = b + ∑ j w j x j + ∑ j < k w j k x j x k + ∑ j < k < l w j k l x j x k x l + . . . {\displaystyle S'({\vec {x}})=b+\sum _{j}w_{j}x_{j}+\sum _{j In computational geometry, the farthest-first traversal of a compact metric space is a sequence of points in the space, where the first point is selected arbitrarily and each successive point is as far as possible from the set of previously-selected points. The same concept can also be applied to a finite set of geometric points, by restricting the selected points to belong to the set or equivalently by considering the finite metric space generated by these points. For a finite metric space or finite set of geometric points, the resulting sequence forms a permutation of the points, also known as the greedy permutation. Every prefix of a farthest-first traversal provides a set of points that is widely spaced and close to all remaining points. More precisely, no other set of equally many points can be spaced more than twice as widely, and no other set of equally many points can be less than half as far to its farthest remaining point. In part because of these properties, farthest-point traversals have many applications, including the approximation of the traveling salesman problem and the metric k-center problem. They may be constructed in polynomial time, or (for low-dimensional Euclidean spaces) approximated in near-linear time. == Definition and properties == A farthest-first traversal is a sequence of points in a compact metric space, with each point appearing at most once. If the space is finite, each point appears exactly once, and the traversal is a permutation of all of the points in the space. The first point of the sequence may be any point in the space. Each point p after the first must have the maximum possible distance to the set of points earlier than p in the sequence, where the distance from a point to a set is defined as the minimum of the pairwise distances to points in the set. A given space may have many different farthest-first traversals, depending both on the choice of the first point in the sequence (which may be any point in the space) and on ties for the maximum distance among later choices. Farthest-point traversals may be characterized by the following properties. Fix a number k, and consider the prefix formed by the first k points of the farthest-first traversal of any metric space. Let r be the distance between the final point of the prefix and the other points in the prefix. Then this subset has the following two properties: All pairs of the selected points are at distance at least r from each other, and All points of the metric space are at distance at most r from the subset. Conversely any sequence having these properties, for all choices of k, must be a farthest-first traversal. These are the two defining properties of a Delone set, so each prefix of the farthest-first traversal forms a Delone set. == Applications == Rosenkrantz, Stearns & Lewis (1977) used the farthest-first traversal to define the farthest-insertion heuristic for the travelling salesman problem. This heuristic finds approximate solutions to the travelling salesman problem by building up a tour on a subset of points, adding one point at a time to the tour in the ordering given by a farthest-first traversal. To add each point to the tour, one edge of the previous tour is broken and replaced by a pair of edges through the added point, in the cheapest possible way. Although Rosenkrantz et al. prove only a logarithmic approximation ratio for this method, they show that in practice it often works better than other insertion methods with better provable approximation ratios. Later, the same sequence of points was popularized by Gonzalez (1985), who used it as part of greedy approximation algorithms for two problems in clustering, in which the goal is to partition a set of points into k clusters. One of the two problems that Gonzalez solve in this way seeks to minimize the maximum diameter of a cluster, while the other, known as the metric k-center problem, seeks to minimize the maximum radius, the distance from a chosen central point of a cluster to the farthest point from it in the same cluster. For instance, the k-center problem can be used to model the placement of fire stations within a city, in order to ensure that every address within the city can be reached quickly by a fire truck. For both clustering problems, Gonzalez chooses a set of k cluster centers by selecting the first k points of a farthest-first traversal, and then creates clusters by assigning each input point to the nearest cluster center. If r is the distance from the set of k selected centers to the next point at position k + 1 in the traversal, then with this clustering every point is within distance r of its center and every cluster has diameter at most 2r. However, the subset of k centers together with the next point are all at distance at least r from each other, and any k-clustering would put some two of these points into a single cluster, with one of them at distance at least r/2 from its center and with diameter at least r. Thus, Gonzalez's heuristic gives an approximation ratio of 2 for both clustering problems. Gonzalez's heuristic was independently rediscovered for the metric k-center problem by Dyer & Frieze (1985), who applied it more generally to weighted k-center problems. Another paper on the k-center problem from the same time, Hochbaum & Shmoys (1985), achieves the same approximation ratio of 2, but its techniques are different. Nevertheless, Gonzalez's heuristic, and the name "farthest-first traversal", are often incorrectly attributed to Hochbaum and Shmoys. For both the min-max diameter clustering problem and the metric k-center problem, these approximations are optimal: the existence of a polynomial-time heuristic with any constant approximation ratio less than 2 would imply that P = NP. As well as for clustering, the farthest-first traversal can also be used in another type of facility location problem, the max-min facility dispersion problem, in which the goal is to choose the locations of k different facilities so that they are as far apart from each other as possible. More precisely, the goal in this problem is to choose k points from a given metric space or a given set of candidate points, in such a way as to maximize the minimum pairwise distance between the selected points. Again, this can be approximated by choosing the first k points of a farthest-first traversal. If r denotes the distance of the kth point from all previous points, then every point of the metric space or the candidate set is within distance r of the first k − 1 points. By the pigeonhole principle, some two points of the optimal solution (whatever it is) must both be within distance r of the same point among these first k − 1 chosen points, and (by the triangle inequality) within distance 2r of each other. Therefore, the heuristic solution given by the farthest-first traversal is within a factor of two of optimal. Other applications of the farthest-first traversal include color quantization (clustering the colors in an image to a smaller set of representative colors), progressive scanning of images (choosing an order to display the pixels of an image so that prefixes of the ordering produce good lower-resolution versions of the whole image rather than filling in the image from top to bottom), point selection in the probabilistic roadmap method for motion planning, simplification of point clouds, generating masks for halftone images, hierarchical clustering, finding the similarities between polygon meshes of similar surfaces, choosing diverse and high-value observation targets for underwater robot exploration, fault detection in sensor networks, modeling phylogenetic diversity, matching vehicles in a heterogenous fleet to customer delivery requests, uniform distribution of geodetic observatories on the Earth's surface or of other types of sensor network, generation of virtual point lights in the instant radiosity computer graphics rendering method, and geometric range searching data structures. == Algorithms == === Greedy exact algorithm === The farthest-first traversal of a finite point set may be computed by a greedy algorithm that maintains the distance of each point from the previously selected points, performing the following steps: Initialize the sequence of selected points to the empty sequence, and the distances of each point to the selected points to infinity. While not all points have been selected, repeat the following steps: Scan the list of not-yet-selected points to find a point p that has the maximum distance from the selected points. Remove p from the not-yet-selected points and add it to the end of the sequence of selected points. For each remaining not-yet-selected point q, replace the distance stored for q by the minimum of its old value and the distance from p to q. For a set of n points, this algorithm takes O(n2) steps and O(n2) distance computations. === Approximations === A faster approximation algorithm, given by Har-Peled & Mendel (2006), applie Computational humor is a branch of computational linguistics and artificial intelligence which uses computers in humor research. It is a relatively new area, with the first dedicated conference organized in 1996. The first "computer model of a sense of humor" was suggested by Suslov as early as 1992. Investigation of the general scheme of the information processing show a possibility of a specific malfunction, conditioned by the necessity of a quick deletion from consciousness of a false version. This specific malfunction can be identified with a humorous effect on the psychological grounds; however, an essentially new ingredient, a role of timing, is added to a well known role of ambiguity. In biological systems, a sense of humour inevitably develops in the course of evolution, because its biological function consists in quickening the transmission of processed information into consciousness and in a more effective use of brain resources. A realization of this algorithm in neural networks explains naturally the mechanism of laughter: deletion of a false version corresponds to zeroing of some part of the neural network and excessive energy of neurons is thrown out to the motor cortex, arousing muscular contractions. Unfortunately, a practical realization of this algorithm needs extensive databases, whose creation in the automatic regime was suggested only recently . As a result, this magistral direction was not developed properly and subsequent investigations (see below) accepted somewhat specialized colouring. == Joke generators == === Pun generation === An approach to analysis of humor is classification of jokes. A further step is an attempt to generate jokes basing on the rules that underlie classification. Simple prototypes for computer pun generation were reported in the early 1990s, based on a natural language generator program, VINCI. Graeme Ritchie and Kim Binsted in their 1994 research paper described a computer program, JAPE, designed to generate question-answer-type puns from a general, i.e., non-humorous, lexicon. (The program name is an acronym for "Joke Analysis and Production Engine".) Some examples produced by JAPE are: Q: What is the difference between leaves and a car? A: One you brush and rake, the other you rush and brake. Q: What do you call a strange market? A: A bizarre bazaar. Since then the approach has been improved, and the latest report, dated 2007, describes the STANDUP joke generator, implemented in the Java programming language. The STANDUP generator was tested on children within the framework of analyzing its usability for language skills development for children with communication disabilities, e.g., because of cerebral palsy. (The project name is an acronym for "System To Augment Non-speakers' Dialog Using Puns" and an allusion to standup comedy.) Children responded to this "language playground" with enthusiasm, and showed marked improvement on certain types of language tests. The two young people, who used the system over a ten-week period, regaled their peers, staff, family and neighbors with jokes such as: "What do you call a spicy missile? A hot shot!" Their joy and enthusiasm at entertaining others was inspirational. === Other === Stock and Strapparava described a program to generate funny acronyms. == Joke recognition == A statistical machine learning algorithm to detect whether a sentence contained a "That's what she said" double entendre was developed by Kiddon and Brun (2011). There is an open-source Python implementation of Kiddon & Brun's TWSS system. A program to recognize knock-knock jokes was reported by Taylor and Mazlack. This kind of research is important in analysis of human–computer interaction. An application of machine learning techniques for the distinguishing of joke texts from non-jokes was described by Mihalcea and Strapparava (2006). Takizawa et al. (1996) reported on a heuristic program for detecting puns in the Japanese language. == Applications == A possible application for assistance in language acquisition is described in the section "Pun generation". Another envisioned use of joke generators is in cases of a steady supply of jokes where quantity is more important than quality. Another obvious, yet remote, direction is automated joke appreciation. It is known that humans interact with computers in ways similar to interacting with other humans that may be described in terms of personality, politeness, flattery, and in-group favoritism. Therefore, the role of humor in human–computer interaction is being investigated. In particular, humor generation in user interface to ease communications with computers was suggested. Craig McDonough implemented the Mnemonic Sentence Generator, which converts passwords into humorous sentences. Based on the incongruity theory of humor, it is suggested that the resulting meaningless but funny sentences are easier to remember. For example, the password AjQA3Jtv is converted into "Arafat joined Quayle's Ant, while TARAR Jeopardized thurmond's vase," an example chosen by combining politicians names with verbs and common nouns. == Related research == John Allen Paulos is known for his interest in mathematical foundations of humor. His book Mathematics and Humor: A Study of the Logic of Humor demonstrates structures common to humor and formal sciences (mathematics, linguistics) and develops a mathematical model of jokes based on catastrophe theory. Conversational systems which have been designed to take part in Turing test competitions generally have the ability to learn humorous anecdotes and jokes. Because many people regard humor as something particular to humans, its appearance in conversation can be quite useful in convincing a human interrogator that a hidden entity, which could be a machine or a human, is in fact a human. In statistics, multinomial logistic regression is a classification method that generalizes logistic regression to multiclass problems, i.e. with more than two possible discrete outcomes. That is, it is a model that is used to predict the probabilities of the different possible outcomes of a categorically distributed dependent variable, given a set of independent variables (which may be real-valued, binary-valued, categorical-valued, etc.). Multinomial logistic regression is known by a variety of other names, including polytomous LR, multiclass LR, softmax regression, multinomial logit (mlogit), the maximum entropy (MaxEnt) classifier, and the conditional maximum entropy model. == Background == Multinomial logistic regression is used when the dependent variable in question is nominal (equivalently categorical, meaning that it falls into any one of a set of categories that cannot be ordered in any meaningful way) and for which there are more than two categories. Some examples would be: Which major will a college student choose, given their grades, stated likes and dislikes, etc.? Which blood type does a person have, given the results of various diagnostic tests? In a hands-free mobile phone dialing application, which person's name was spoken, given various properties of the speech signal? Which candidate will a person vote for, given particular demographic characteristics? Which country will a firm locate an office in, given the characteristics of the firm and of the various candidate countries? These are all statistical classification problems. They all have in common a dependent variable to be predicted that comes from one of a limited set of items that cannot be meaningfully ordered, as well as a set of independent variables (also known as features, explanators, etc.), which are used to predict the dependent variable. Multinomial logistic regression is a particular solution to classification problems that use a linear combination of the observed features and some problem-specific parameters to estimate the probability of each particular value of the dependent variable. The best values of the parameters for a given problem are usually determined from some training data (e.g. some people for whom both the diagnostic test results and blood types are known, or some examples of known words being spoken). == Assumptions == The multinomial logistic model assumes that data are case-specific; that is, each independent variable has a single value for each case. As with other types of regression, there is no need for the independent variables to be statistically independent from each other (unlike, for example, in a naive Bayes classifier); however, collinearity is assumed to be relatively low, as it becomes difficult to differentiate between the impact of several variables if this is not the case. If the multinomial logit is used to model choices, it relies on the assumption of independence of irrelevant alternatives (IIA), which is not always desirable. This assumption states that the odds of preferring one class over another do not depend on the presence or absence of other "irrelevant" alternatives. For example, the relative probabilities of taking a car or bus to work do not change if a bicycle is added as an additional possibility. This allows the choice of K alternatives to be modeled as a set of K − 1 independent binary choices, in which one alternative is chosen as a "pivot" and the other K − 1 compared against it, one at a time. The IIA hypothesis is a core hypothesis in rational choice theory; however numerous studies in psychology show that individuals often violate this assumption when making choices. An example of a problem case arises if choices include a car and a blue bus. Suppose the odds ratio between the two is 1 : 1. Now if the option of a red bus is introduced, a person may be indifferent between a red and a blue bus, and hence may exhibit a car : blue bus : red bus odds ratio of 1 : 0.5 : 0.5, thus maintaining a 1 : 1 ratio of car : any bus while adopting a changed car : blue bus ratio of 1 : 0.5. Here the red bus option was not in fact irrelevant, because a red bus was a perfect substitute for a blue bus. If the multinomial logit is used to model choices, it may in some situations impose too much constraint on the relative preferences between the different alternatives. It is especially important to take into account if the analysis aims to predict how choices would change if one alternative were to disappear (for instance if one political candidate withdraws from a three candidate race). Other models like the nested logit or the multinomial probit may be used in such cases as they allow for violation of the IIA. == Model == === Introduction === There are multiple equivalent ways to describe the mathematical model underlying multinomial logistic regression. This can make it difficult to compare different treatments of the subject in different texts. The article on logistic regression presents a number of equivalent formulations of simple logistic regression, and many of these have analogues in the multinomial logit model. The idea behind all of them, as in many other statistical classification techniques, is to construct a linear predictor function that constructs a score from a set of weights that are linearly combined with the explanatory variables (features) of a given observation using a dot product: score ( X i , k ) = β k ⋅ X i , {\displaystyle \operatorname {score} (\mathbf {X} _{i},k)={\boldsymbol {\beta }}_{k}\cdot \mathbf {X} _{i},} where Xi is the vector of explanatory variables describing observation i, βk is a vector of weights (or regression coefficients) corresponding to outcome k, and score(Xi, k) is the score associated with assigning observation i to category k. In discrete choice theory, where observations represent people and outcomes represent choices, the score is considered the utility associated with person i choosing outcome k. The predicted outcome is the one with the highest score. The difference between the multinomial logit model and numerous other methods, models, algorithms, etc. with the same basic setup (the perceptron algorithm, support vector machines, linear discriminant analysis, etc.) is the procedure for determining (training) the optimal weights/coefficients and the way that the score is interpreted. In particular, in the multinomial logit model, the score can directly be converted to a probability value, indicating the probability of observation i choosing outcome k given the measured characteristics of the observation. This provides a principled way of incorporating the prediction of a particular multinomial logit model into a larger procedure that may involve multiple such predictions, each with a possibility of error. Without such means of combining predictions, errors tend to multiply. For example, imagine a large predictive model that is broken down into a series of submodels where the prediction of a given submodel is used as the input of another submodel, and that prediction is in turn used as the input into a third submodel, etc. If each submodel has 90% accuracy in its predictions, and there are five submodels in series, then the overall model has only 0.95 = 59% accuracy. If each submodel has 80% accuracy, then overall accuracy drops to 0.85 = 33% accuracy. This issue is known as error propagation and is a serious problem in real-world predictive models, which are usually composed of numerous parts. Predicting probabilities of each possible outcome, rather than simply making a single optimal prediction, is one means of alleviating this issue. === Setup === The basic setup is the same as in logistic regression, the only difference being that the dependent variables are categorical rather than binary, i.e. there are K possible outcomes rather than just two. The following description is somewhat shortened; for more details, consult the logistic regression article. ==== Data points ==== Specifically, it is assumed that we have a series of N observed data points. Each data point i (ranging from 1 to N) consists of a set of M explanatory variables x1,i ... xM,i (also known as independent variables, predictor variables, features, etc.), and an associated categorical outcome Yi (also known as dependent variable, response variable), which can take on one of K possible values. These possible values represent logically separate categories (e.g. different political parties, blood types, etc.), and are often described mathematically by arbitrarily assigning each a number from 1 to K. The explanatory variables and outcome represent observed properties of the data points, and are often thought of as originating in the observations of N "experiments" — although an "experiment" may consist of nothing more than gathering data. The goal of multinomial logistic regression is to construct a model that explains the relationship between the explanatory variables and the outcome, so thaFarthest-first traversal
Computational humor
Multinomial logistic regression