Meta tags:

Headings (most frequently used words):

lightgbm, contents, overview, gradient, based, one, side, sampling, exclusive, feature, bundling, see, also, references, further, reading, external, links,

Text of the page (most frequently used words):
#lightgbm (40), the (33), microsoft (31), learning (24), and (22), net (20), machine (17), windows (15), gradient (15), software (14), open (14), for (12), edit (12), data (10), framework (10), source (9), with (8), deep (8), github (8), features (8), based (8), exclusive (8), feature (8), one (8), wikipedia (7), foundation (7), license (7), free (7), other (7), boosting (7), this (6), python (6), research (6), neural (6), labs (6), tools (6), asp (6), 2020 (6), decision (6), tree (6), which (6), search (5), page (5), toolkit (5), xgboost (5), learn (5), languages (4), contents (4), may (4), using (4), algorithms (4), scikit (4), com (4), prover (4), main (4), linux (4), library (4), file (4), system (4), overview (4), links (4), guolin (4), bundling (4), side (4), sampling (4), model (4), valley (4), algorithm (4), hide (4), move (4), sidebar (4), view (3), about (3), parameters (3), short (3), different (3), from (3), 2016 (3), mit (3), intelligence (3), list (3), comparison (3), see (3), also (3), cognitive (3), deepspeed (3), studio (3), apache (3), spark (3), infer (3), catboost (3), live (3), operating (3), systems (3), network (3), projects (3), conference (3), distributed (3), project (3), visual (3), basic (3), information (3), highly (3), via (3), july (3), readthedocs (3), documentation (3), cite (3), 2024 (3), efb (3), method (3), are (3), level (3), does (3), not (3), informative (3), instead (3), trees (3), used (3), toggle (2), table (2), developers (2), code (2), contact (2), privacy (2), policy (2), text (2), available (2), under (2), terms (2), use (2), was (2), october (2), categories (2), deprecated (2), description (2), wikidata (2), programmed (2), mining (2), applied (2), portal (2), pytorch (2), keras (2), vowpal (2), wabbit (2), online (2), category (2), docs (2), current (2), msr (2), allegiance (2), detours (2), comic (2), chat (2), mimalloc (2), dryad (2), barrelfish (2), theorem (2), worldwide (2), telescope (2), temporal (2), seal (2), onefuzz (2), automatic (2), graph (2), layout (2), orleans (2), verona (2), lean (2), bosque (2), related (2), template (2), forms (2), frameworks (2), sdk (2), xml (2), design (2), language (2), development (2), package (2), manager (2), external (2), 978 (2), isbn (2), more (2), meng (2), thomas (2), taifeng (2), wang (2), wei (2), chen (2), weidong (2), qiwei (2), 2017 (2), processing (2), pdf (2), efficient (2), tie (2), yan (2), liu (2), further (2), reading (2), 2022 (2), uses (2), help (2), citeseerx (2), nov (2), how (2), them (2), early (2), type (2), references (2), sparse (2), space (2), many (2), values (2), efficiency (2), while (2), maintaining (2), high (2), accuracy (2), bundle (2), called (2), goss (2), developed (2), that (2), equally (2), slopes (2), descent (2), point (2), most (2), supports (2), macos (2), including (2), advantages (2), loss (2), two (2), grow (2), wise (2), row (2), implementations (2), leaf (2), sorted (2), years (2), ago (2), release (2), appearance (2), upload (2), changes (2), history (2), read (2), article (2), log (2), create (2), account (2), donate (2), menu (2), add, topic, mobile, cookie, statement, statistics, conduct, legal, safety, contacts, disclaimers, additional, apply, site, you, agree, registered, trademark, non, profit, organization, wikimedia, inc, creative, commons, attribution, sharealike, rendered, parsoid, last, edited, 2026, utc, hidden, cs1, errors, articles, artificial, retrieved, https, org, index, php, title, oldid, 1378514591, datasets, torch, theano, tensorrt, tensorflow, lightning, plaidml, openvino, opennn, designer, mxnet, mindspore, jax, hugging, face, transformers, horovod, flux, fastai, deeplearning4j, chainer, caffe, bigdl, singa, wolfram, mathematica, weka, mllib, shogun, multiflow, rapidminer, pymc, orange, mlx, apple, mlpack, matlab, massive, analysis, libsvm, knime, jubatus, h2o, gensim, fasttext, elki, dlib, systemds, mahout, accord, office, adcenter, academic, kodu, fuse, photosynth, clipboard, listas, deepfish, discontinued, zoom, seadragon, pivot, kinect, xbox, songsmith, trueskill, group, shot, cleartype, sensecam, touchlight, pixelsense, sideshow, launched, products, joins, accelerator, apis, verve, singularity, midori, homeos, slam, photodna, nodexl, mylifebits, livestation, image, composite, editor, illumiroom, holoportation, digits, wallop, penny, black, honeymonkey, gazelle, avalanche, ajaxview, networking, internet, confidential, consortium, bitvault, grid, computing, secpal, polyphonic, phoenix, bartok, compilers, outercurve, security, letter, hobbyists, specification, promise, codeplex, forges, reciprocal, public, licenses, subsystem, tpm, reference, implementation, tlaps, silk, reactivex, gollum, fourq, flexwiki, extensible, storage, engine, chronozoom, azure, sonic, dos, xunit, xsp, xdp, wix, winobjc, winjs, presentation, umdf, kmdf, driver, communication, app, vott, voldemort, virtual, git, vcpkg, prove, tla, toolbox, transformation, svnbridge, stylecop, signalr, sandcastle, roslyn, recursiveextractor, procmon, procdump, playwright, service, mesh, exchange, management, infrastructure, nuget, npm, msquic, msbuild, monodevelop, mono, mod_mono, mixed, reality, enterprise, standard, managed, extensibility, fluid, fluent, entity, electron, ebpf, dynamic, runtime, diskspd, dapr, clr, profiler, chakracore, ccf, winrt, blazor, bitfunnel, babylon, avalonia, web, razor, mvc, core, ajax, airsim, micro, maui, gadgeteer, typescript, small, powershell, power, ironruby, ironpython, dafny, programming, zork, video, games, notepad, console, calculator, terminal, powertoys, writer, family, show, atom, movie, maker, applications, shared, initiative, foss, van, wyk, andrich, 2023, 1800564749, packt, publishing, quinto, butch, 4842, 5668, apress, next, generation, covers, nlp, finely, trainers, namespace, pypi, finley, advances, parameter, journal, manish, mehta, rakesh, agrawal, jorma, rissanen, sliq, fast, scalable, classifier, 7734, international, extending, database, technology, official, boosters, shallow, 2018, avanwyk, neptune, understanding, tune, kopitar, leon, kocbek, primoz, cilar, leona, sheikh, aziz, stiglic, gregor, 11981, www, nature, 32686721, pmid, 7371679, pmc, 1038, s41598, 020, 68771, doi, 2020natsr, 1011981k, bibcode, scientific, reports, detection, diabetes, mellitus, prediction, models, brownlee, jason, march, releases, binning, tabpfn, near, lossless, reduce, number, effective, nearly, implying, they, rarely, take, nonzero, simultaneously, hot, encoded, perfect, example, bundles, these, reducing, dimensionality, improve, into, single, contrast, first, rely, assumption, all, treats, points, smaller, gradients, shallower, less, randomly, dropping, intended, filter, out, have, been, influenced, noise, allowing, accurately, underlying, relationships, boosted, typically, whole, set, calculate, however, commonly, assumes, every, when, thinks, possible, configurations, lowest, part, closely, fits, metaphor, walks, directions, much, lower, becomes, works, licensed, gbt, has, optimization, parallel, training, multiple, functions, regularization, bagging, stopping, major, difference, between, lies, construction, grows, will, choose, max, delta, besides, widely, searches, best, split, implements, optimized, histogram, yields, great, both, memory, consumption, utilizes, novel, techniques, allow, run, faster, mart, gbm, gbrt, gbdt, originally, tasks, focus, performance, scalability, classification, ranking, light, repository, website, written, january, stable, contributors, original, author, encyclopedia, item, printable, version, download, print, export, switch, legacy, parser, get, shortened, url, permanent, link, what, here, general, actions, english, talk, 한국어, 日本語, فارسی, বাংলা, top, personal, special, pages, recent, community, contribute, random, events, navigation, jump, content,

Text of the page (random words):
lightgbm wikipedia jump to content main menu main menu move to sidebar hide navigation main page contents current events random article about wikipedia contact us contribute help learn to edit community portal recent changes upload file special pages search search appearance donate create account log in personal tools donate create account log in contents move to sidebar hide top 1 overview 2 gradient based one side sampling 3 exclusive feature bundling 4 see also 5 references 6 further reading 7 external links toggle the table of contents lightgbm 4 languages বাংলা فارسی 日本語 한국어 edit links article talk english read edit view history tools tools move to sidebar hide actions read edit view history general what links here related changes upload file permanent link page information cite this page get shortened url switch to legacy parser print export download as pdf printable version in other projects wikidata item appearance move to sidebar hide from wikipedia the free encyclopedia microsoft open source gradient boosting framework for machine learning lightgbm original author guolin ke 1 microsoft research developers microsoft and lightgbm contributors 2 release 2016 10 years ago 2016 stable release v4 3 0 3 january 15 2024 2 years ago 2024 01 15 written in c python r c operating system windows macos linux type machine learning gradient boosting framework license mit license website lightgbm readthedocs io repository github com microsoft lightgbm lightgbm short for light gradient boosting machine is a free and open source distributed gradient boosting framework for machine learning originally developed by microsoft 4 5 it is based on decision tree algorithms and used for ranking classification and other machine learning tasks the development focus is on performance and scalability overview edit the lightgbm framework supports different algorithms including gbt gbdt gbrt gbm mart 6 7 and rf 8 lightgbm has many of xgboost s advantages including sparse optimization parallel training multiple loss functions regularization bagging and early stopping a major difference between the two lies in the construction of trees lightgbm does not grow a tree level wise row by row as most other implementations do 9 instead it grows trees leaf wise it will choose the leaf with max delta loss to grow 10 besides lightgbm does not use the widely used sorted based decision tree learning algorithm which searches the best split point on sorted feature values 11 as xgboost or other implementations do instead lightgbm implements a highly optimized histogram based decision tree learning algorithm which yields great advantages on both efficiency and memory consumption 12 the lightgbm algorithm utilizes two novel techniques called gradient based one side sampling goss and exclusive feature bundling efb which allow the algorithm to run faster while maintaining a high level of accuracy 13 lightgbm works on linux windows and macos and supports c python 14 r and c 15 the source code is licensed under mit license and available on github 16 gradient based one side sampling edit when using gradient descent one thinks about the space of possible configurations of the model as a valley in which the lowest part of the valley is the model which most closely fits the data in this metaphor one walks in different directions to learn how much lower the valley becomes typically in gradient descent one uses the whole set of data to calculate the valley s slopes however this commonly used method assumes that every data point is equally informative by contrast gradient based one side sampling goss a method first developed for gradient boosted decision trees does not rely on the assumption that all data are equally informative instead it treats data points with smaller gradients shallower slopes as less informative by randomly dropping them this is intended to filter out data which may have been influenced by noise allowing the model to more accurately model the underlying relationships in the data 13 exclusive feature bundling edit exclusive feature bundling efb is a near lossless method to reduce the number of effective features in a sparse feature space many features are nearly exclusive implying they rarely take nonzero values simultaneously one hot encoded features are a perfect example of exclusive features efb bundles these features reducing dimensionality to improve efficiency while maintaining a high level of accuracy the bundle of exclusive features into a single feature is called an exclusive feature bundle 13 see also edit tabpfn ml net data binning catboost scikit learn comparison of machine learning software references edit guolin ke github microsoft lightgbm github 7 july 2022 releases microsoft lightgbm github brownlee jason march 31 2020 gradient boosting with scikit learn xgboost lightgbm and catboost kopitar leon kocbek primoz cilar leona sheikh aziz stiglic gregor july 20 2020 early detection of type 2 diabetes mellitus using machine learning based prediction models scientific reports 10 1 11981 bibcode 2020natsr 1011981k doi 10 1038 s41598 020 68771 z pmc 7371679 pmid 32686721 via www nature com understanding lightgbm parameters and how to tune them neptune ai may 6 2020 an overview of lightgbm avanwyk may 16 2018 parameters lightgbm 3 0 0 99 documentation lightgbm readthedocs io the gradient boosters iv lightgbm deep shallow features lightgbm official documentation nov 3 2024 manish mehta rakesh agrawal jorma rissanen nov 24 2020 sliq a fast scalable classifier for data mining international conference on extending database technology 18 32 citeseerx 10 1 1 89 7734 cite journal cite uses deprecated parameter citeseerx help features lightgbm 3 1 0 99 documentation lightgbm readthedocs io 1 2 3 ke guolin meng qi finley thomas wang taifeng chen wei ma weidong ye qiwei liu tie yan 2017 lightgbm a highly efficient gradient boosting decision tree advances in neural information processing systems 30 lightgbm lightgbm python package 7 july 2022 via pypi microsoft ml trainers lightgbm namespace docs microsoft com microsoft lightgbm october 6 2020 via github further reading edit guolin ke qi meng thomas finely taifeng wang wei chen weidong ma qiwei ye tie yan liu 2017 lightgbm a highly efficient gradient boosting decision tree pdf neural information processing system quinto butch 2020 next generation machine learning with spark covers xgboost lightgbm spark nlp distributed deep learning with keras and more apress isbn 978 1 4842 5668 8 van wyk andrich 2023 machine learning with lightgbm and python packt publishing isbn 978 1800564749 external links edit github microsoft lightgbm lightgbm microsoft research v t e microsoft free and open source software foss overview microsoft and open source shared source initiative software applications 3d movie maker atom conference xp family show file manager open live writer microsoft comic chat microsoft edit microsoft powertoys terminal windows calculator windows console windows package manager worldwide telescope xml notepad video games allegiance zork programming languages bosque c dafny f f gw basic ironpython ironruby lean p power fx powershell project verona q small basic online typescript visual basic frameworks development tools net net framework net gadgeteer net maui net micro framework airsim asp net asp net ajax asp net core asp net mvc asp net razor asp net web forms avalonia babylon js bitfunnel blazor c winrt ccf chakracore clr profiler dapr deepspeed diskspd dryad dynamic language runtime ebpf on windows electron entity framework fluent design system fluid framework infer net lightgbm managed extensibility framework microsoft automatic graph layout microsoft c standard library microsoft cognitive toolkit microsoft design language microsoft detours microsoft enterprise library microsoft seal mimalloc mixed reality toolkit ml net mod_mono mono monodevelop msbuild msquic neural network intelligence npm nuget onefuzz open management infrastructure open neural network exchange open service mesh open xml sdk orleans playwright procdump procmon python tools for visual studio r tools for visual studio recursiveextractor roslyn sandcastle signalr stylecop svnbridge t2 temporal prover text template transformation toolkit tla toolbox u prove vcpkg virtual file system for git voldemort vott vowpal wabbit windows app sdk windows communication foundation windows driver frameworks kmdf umdf windows forms windows presentation foundation windows template library windows ui library winjs winobjc wix xdp for windows xsp xunit net z3 theorem prover operating systems ms dos v1 25 v2 0 v4 0 barrelfish sonic azure linux other chronozoom extensible storage engine flexwiki fourq gollum project mu reactivex silk tlaps tpm 2 0 reference implementation windows subsystem for linux licenses microsoft public license microsoft reciprocal license forges codeplex github related net foundation f software foundation microsoft open specification promise open letter to hobbyists open source security foundation outercurve foundation category v t e microsoft research msr main projects languages compilers bartok bosque cω f lean p project verona phoenix polyphonic c secpal distributed grid computing bitvault confidential consortium framework deepspeed orleans internet networking ajaxview avalanche conference xp gazelle honeymonkey penny black wallop other projects automatic graph layout cognitive toolkit digits holoportation illumiroom image composite editor infer net lightgbm livestation mylifebits neural network intelligence nodexl onefuzz photodna seal slam t2 temporal prover worldwide telescope z3 theorem prover operating systems barrelfish homeos midori singularity verve apis accelerator dryad joins mimalloc launched as products c comic chat detours f sideshow pixelsense touchlight sensecam cleartype group shot allegiance trueskill songsmith xbox kinect msr labs applied research live labs current pivot seadragon deep zoom discontinued deepfish listas live clipboard photosynth fuse labs docs com kodu other labs academic search adcenter labs office labs category v t e machine learning software machine learning accord net apache mahout apache systemds catboost dlib elki fasttext gensim h2o infer net jubatus knime libsvm lightgbm massive online analysis matlab ml net mlpack mlx apple orange pymc rapidminer scikit learn scikit multiflow shogun spark mllib vowpal wabbit weka wolfram mathematica xgboost deep learning apache singa bigdl caffe chainer deeplearning4j deep learning studio deepspeed fastai flux horovod hugging face transformers jax keras microsoft cognitive toolkit mindspore mxnet neural designer opennn openvino plaidml pytorch pytorch lightning tensorflow tensorrt theano torch see also comparison of deep learning software comparison of machine learning software list of datasets for machine learning research list of machine learning algorithms list of deep learning algorithms machine learning deep learning portal free and open source software retrieved from https en wikipedia org w index php title lightgbm oldid 1378514591 categories applied machine learning data mining and machine learning software free software programmed in c microsoft free software microsoft research open source artificial intelligence software using the mit license 2016 software free software programmed in python hidden categories articles with short description short description is different from wikidata cs1 errors deprecated parameters this page was last edited on 4 october 2026 at 21 08 utc page was rendered with parsoid text is available under the creative commons attribution sharealike 4 0 license additional terms may apply by using this site you agree to the terms of use and privacy policy wikipedia is a registered trademark of the wikimedia foundation inc a non profit organization privacy policy about wikipedia disclaimers contact wikipedia legal safety contacts code of conduct developers statistics cookie statement mobile view search search toggle the table of contents lightgbm 4 languages add topic