{"id":53278,"date":"2019-11-28T00:00:00","date_gmt":"2019-11-27T21:00:00","guid":{"rendered":"https:\/\/prohoster.info\/blog\/blog_prohoster\/9-podhodov-dlya-vyyavleniya-anomalij"},"modified":"2020-02-18T14:01:10","modified_gmt":"2020-02-18T11:01:10","slug":"9-podhodov-dlya-vyyavleniya-anomalij","status":"publish","type":"post","link":"https:\/\/prohoster.info\/sv\/blog\/news\/9-podhodov-dlya-vyyavleniya-anomalij","title":{"rendered":"9 metoder f\u00f6r att identifiera anomalier","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

I den f\u00f6reg\u00e5ende artikeln<\/a><\/noindex> Vi har pratat om prognoser f\u00f6r tidsserier. En logisk forts\u00e4ttning skulle vara en artikel om att uppt\u00e4cka avvikelser. <\/p>\n

Till\u00e4mpning<\/h3>\n

\nAvvikelseuppt\u00e4ckning anv\u00e4nds inom omr\u00e5den som: <\/p>\n

1) F\u00f6ruts\u00e4gelse av maskinfel<\/h3>\n

\nTill exempel, 2010 attackerades iranska centrifuger av viruset Stuxnet, vilket st\u00e4llde in utrustningen p\u00e5 en suboptimal drift och orsakade att en del av utrustningen skadades p\u00e5 grund av \u00f6kat slitage. <\/p>\n

Om man hade anv\u00e4nt algoritmer f\u00f6r avvikelseuppt\u00e4ckning p\u00e5 utrustningen, skulle s\u00e5dana driftstopp ha kunnat undvikas. <\/p>\n

\"9
\n<\/a><\/noindex>
\nAvvikelseuppt\u00e4ckning i maskinv\u00e5rd anv\u00e4nds inte bara inom k\u00e4rnkraftindustrin, utan \u00e4ven inom metallurgi och drift av flygmotorer. Och inom andra omr\u00e5den d\u00e4r anv\u00e4ndning av prediktiv diagnos \u00e4r billigare \u00e4n potentiella f\u00f6rluster vid of\u00f6rutsedda fel.<\/p>\n

2) F\u00f6ruts\u00e4gelse av bedr\u00e4gliga handlingar<\/h3>\n

\nOm pengar tas fr\u00e5n ett kort som du anv\u00e4nder i Podolsk, i Albanien, kan det vara bra att dubbelkolla transaktionerna.<\/p>\n

3) Uppt\u00e4ckte avvikande kundm\u00f6nster<\/h3>\n

\nOm en del av kunderna visar avvikande beteende kan det finnas ett problem som du inte k\u00e4nner till.<\/p>\n

4) Uppt\u00e4ckte avvikande efterfr\u00e5gan och belastning<\/h3>\n

\nOm f\u00f6rs\u00e4ljningen i din FMCG-butik sjunker under det f\u00f6rutsagda konfidensintervallet, \u00e4r det v\u00e4rt att s\u00f6ka efter orsaken.<\/p>\n

Metoder f\u00f6r att uppt\u00e4cka avvikelser<\/h2>\n

<\/p>\n

1) St\u00f6dvektormetoden med en klass One-Class SVM<\/h3>\n

\nDenna metod passar n\u00e4r tr\u00e4ningsdatam\u00e4ngden f\u00f6ljer en normalf\u00f6rdelning och testdatam\u00e4ngden inneh\u00e5ller avvikelser.<\/p>\n

One-Class SVM bygger en icke-linj\u00e4r yta runt origo. En tr\u00f6skel kan s\u00e4ttas f\u00f6r att avg\u00f6ra vilka data som ska betraktas som avvikande.<\/p>\n

Enligt erfarenheten fr\u00e5n v\u00e5rt team DATA4 \u00e4r One-Class SVM den mest anv\u00e4nda algoritmen f\u00f6r att l\u00f6sa avvikelseuppt\u00e4ckningsproblem. <\/p>\n

\"9
\n<\/p>\n

2) Isolerande skog \u2013 isolate forest<\/h3>\n

\nVid \"slumpt\u00e4ckning\" av tr\u00e4d kommer avvikelser att hamna i blad p\u00e5 tidiga stadier (med en liten tr\u00e4ddjup), dvs. avvikelser \u00e4r l\u00e4ttare att \"isolera\". Identifikation av avvikande v\u00e4rden sker tidigt i algoritmens iterationer.<\/p>\n

\"9
\n<\/p>\n

3) Elliptisk kuvert och statistiska metoder<\/h3>\n

\nAnv\u00e4nds n\u00e4r data \u00e4r normalt f\u00f6rdelade. Ju n\u00e4rmare m\u00e4tningen ligger svansen av f\u00f6rdelningarna, desto mer avvikande \u00e4r v\u00e4rdet.<\/p>\n

Denna klass kan \u00e4ven inkludera andra statistiska metoder.<\/p>\n

\"9
\n
\n\"9
\nBild fr\u00e5n dyakonov.org<\/i><\/p>\n

4) Metriska metoder<\/h3>\n

\nBland metoderna finns algoritmer som k n\u00e4rmaste grannar, k n\u00e4rmaste granne, ABOD (angle-based outlier detection) eller LOF (local outlier factor).<\/p>\n

Dessa \u00e4r l\u00e4mpliga om avst\u00e5ndet mellan v\u00e4rden i funktionerna \u00e4r likv\u00e4rdigt eller normalt (f\u00f6r att inte m\u00e4ta pytonormar i papegujor). <\/p>\n

Algoritmen k n\u00e4rmaste grannar antar att normala v\u00e4rden \u00e4r placerade inom ett visst omr\u00e5de i m\u00e5ngdimensionellt rum, och att avst\u00e5ndet till avvikelserna kommer att vara st\u00f6rre \u00e4n till den separerande hyperplanen. <\/p>\n

\"9
\n<\/p>\n

5) Kluster metoder<\/h3>\n

\nK\u00e4rnan i kluster metoder \u00e4r att om ett v\u00e4rde \u00e4r mer \u00e4n ett visst avst\u00e5nd fr\u00e5n klustrets centrum, s\u00e5 kan v\u00e4rdet betraktas som en avvikelse. <\/p>\n

Det viktigaste \u00e4r att anv\u00e4nda en algoritm som korrekt klustrar datan, vilket beror p\u00e5 den specifika uppgiften. <\/p>\n

\"9
\n<\/p>\n

6) Huvudkomponentmetoden<\/h3>\n

\nL\u00e4mplig d\u00e4r riktningar f\u00f6r st\u00f6rsta f\u00f6r\u00e4ndring av varians framh\u00e4vs.<\/p>\n

7) Algoritmer baserade p\u00e5 tidsserief\u00f6ruts\u00e4gelser<\/h3>\n

\nId\u00e9n \u00e4r att om ett v\u00e4rde avviker fr\u00e5n f\u00f6rtroendeintervallet f\u00f6r prognosen, betraktas v\u00e4rdet som en avvikelse. F\u00f6r tidsserieprognoser anv\u00e4nds algoritmer som tredubbel utj\u00e4mning, S(ARIMA), boosting, osv.<\/p>\n

Om algoritmer f\u00f6r tidsserieprognoser har n\u00e4mnts i den f\u00f6reg\u00e5ende artikeln.<\/p>\n

\"9
\n<\/p>\n

8) L\u00e4rande med \u00f6vervakning (regression, klassificering)<\/h3>\n

\nOm datan till\u00e5ter anv\u00e4nder vi algoritmer fr\u00e5n linj\u00e4r regression till \u00e5terkommande n\u00e4tverk. Vi m\u00e4ter skillnaden mellan prognosen och det faktiska v\u00e4rdet, och drar slutsatsen om hur mycket datan avviker fr\u00e5n normen. Det \u00e4r viktigt att algoritmen har tillr\u00e4cklig generaliseringsf\u00f6rm\u00e5ga och att tr\u00e4ningssetet inte inneh\u00e5ller avvikande v\u00e4rden.<\/p>\n

9) Modelltester<\/h3>\n

\nVi n\u00e4rmar oss uppgiften att hitta avvikelser som en uppgift att hitta rekommendationer. Vi faktorisera v\u00e5r funktionsmatris med hj\u00e4lp av SVD eller faktoriseringmaskiner, och v\u00e4rden i den nya matrisen som avviker avsev\u00e4rt fr\u00e5n originalet betraktas som avvikande. <\/p>\n

\"9
\n
\nBild fr\u00e5n dyakonov.org<\/i><\/p>\n

Slutsats<\/h3>\n

\nI den h\u00e4r artikeln har vi granskat de grundl\u00e4ggande metoderna f\u00f6r avvikelsedetektering. <\/p>\n

Att identifiera anomalier kan till stor del beskrivas som en konst. Det finns ingen perfekt algoritm eller metod som l\u00f6ser alla problem. Oftare anv\u00e4nds en kombination av metoder f\u00f6r att l\u00f6sa specifika fall. Anomalidetektion utf\u00f6rs med hj\u00e4lp av enklassmetoder f\u00f6r st\u00f6dvektorer, isolerade skogar, metrisk och klustermetoder, samt genom anv\u00e4ndning av huvudkomponenter och tidsserief\u00f6ruts\u00e4gelser.<\/p>\n

Om du k\u00e4nner till andra metoder, skriv g\u00e4rna om dem i kommentarerna till artikeln.
\n
K\u00e4lla: habr.com<\/a><\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"

\u0412 \u043f\u0440\u0435\u0434\u044b\u0434\u0443\u0449\u0435\u0439 \u0441\u0442\u0430\u0442\u044c\u0435 \u043c\u044b \u0433\u043e\u0432\u043e\u0440\u0438\u043b\u0438 \u043e \u043f\u0440\u043e\u0433\u043d\u043e\u0437\u0438\u0440\u043e\u0432\u0430\u043d\u0438\u0438 \u0432\u0440\u0435\u043c\u0435\u043d\u043d\u044b\u0445 \u0440\u044f\u0434\u043e\u0432. \u041b\u043e\u0433\u0438\u0447\u043d\u044b\u043c \u043f\u0440\u043e\u0434\u043e\u043b\u0436\u0435\u043d\u0438\u0435\u043c \u0431\u0443\u0434\u0435\u0442 \u0441\u0442\u0430\u0442\u044c\u044f \u043e \u0432\u044b\u044f\u0432\u043b\u0435\u043d\u0438\u0438 \u0430\u043d\u043e\u043c\u0430\u043b\u0438\u0439. \u041f\u0440\u0438\u043c\u0435\u043d\u0435\u043d\u0438\u0435 \u0412\u044b\u044f\u0432\u043b\u0435\u043d\u0438\u0435 \u0430\u043d\u043e\u043c\u0430\u043b\u0438\u0439 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u0443\u0435\u0442\u0441\u044f \u0432 \u0442\u0430\u043a\u0438\u0445 \u043e\u0431\u043b\u0430\u0441\u0442\u044f\u0445 \u043a\u0430\u043a: 1) \u041f\u0440\u0435\u0434\u0441\u043a\u0430\u0437\u0430\u043d\u0438\u0435 \u043f\u043e\u043b\u043e\u043c\u043e\u043a \u043e\u0431\u043e\u0440\u0443\u0434\u043e\u0432\u0430\u043d\u0438\u044f \u0422\u0430\u043a, \u0432 2010 \u0433\u043e\u0434\u0443 \u0418\u0440\u0430\u043d\u0441\u043a\u0438\u0435 \u0446\u0435\u043d\u0442\u0440\u0438\u0444\u0443\u0433\u0438 \u0431\u044b\u043b\u0438 \u0430\u0442\u0430\u043a\u043e\u0432\u0430\u043d\u044b \u0432\u0438\u0440\u0443\u0441\u043e\u043c Stuxnet, \u043a\u043e\u0442\u043e\u0440\u044b\u0439 \u0437\u0430\u0434\u0430\u043b \u043d\u0435\u043e\u043f\u0442\u0438\u043c\u0430\u043b\u044c\u043d\u044b\u0439 \u0440\u0435\u0436\u0438\u043c \u0440\u0430\u0431\u043e\u0442\u044b \u043e\u0431\u043e\u0440\u0443\u0434\u043e\u0432\u0430\u043d\u0438\u044f \u0438 \u0432\u044b\u0432\u0435\u043b \u0438\u0437 \u0441\u0442\u0440\u043e\u044f \u0447\u0430\u0441\u0442\u044c \u043e\u0431\u043e\u0440\u0443\u0434\u043e\u0432\u0430\u043d\u0438\u044f \u0437\u0430 \u0441\u0447\u0435\u0442 \u0443\u0441\u043a\u043e\u0440\u0435\u043d\u043d\u043e\u0433\u043e \u0438\u0437\u043d\u043e\u0441\u0430. \u0415\u0441\u043b\u0438 […]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[702],"tags":[],"class_list":["post-53278","post","type-post","status-publish","format-standard","hentry","category-news"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\n","aioseo_head_json":{"title":"\ud83e\udd479 metoder f\u00f6r anomalidetektion | ProHoster","description":"I den f\u00f6reg\u00e5ende artikeln diskuterade vi tidsserief\u00f6ruts\u00e4gelser. En logisk forts\u00e4ttning blir en artikel om anomalidetektion.","canonical_url":"https:\/\/prohoster.info\/sv\/blog\/news\/9-podhodov-dlya-vyyavleniya-anomalij","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"sv_SE","og:site_name":"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b","og:type":"article","og:title":"\ud83e\udd479 \u043f\u043e\u0434\u0445\u043e\u0434\u043e\u0432 \u0434\u043b\u044f \u0432\u044b\u044f\u0432\u043b\u0435\u043d\u0438\u044f \u0430\u043d\u043e\u043c\u0430\u043b\u0438\u0439 | ProHoster","og:description":"\u0412 \u043f\u0440\u0435\u0434\u044b\u0434\u0443\u0449\u0435\u0439 \u0441\u0442\u0430\u0442\u044c\u0435 \u043c\u044b \u0433\u043e\u0432\u043e\u0440\u0438\u043b\u0438 \u043e \u043f\u0440\u043e\u0433\u043d\u043e\u0437\u0438\u0440\u043e\u0432\u0430\u043d\u0438\u0438 \u0432\u0440\u0435\u043c\u0435\u043d\u043d\u044b\u0445 \u0440\u044f\u0434\u043e\u0432. \u041b\u043e\u0433\u0438\u0447\u043d\u044b\u043c \u043f\u0440\u043e\u0434\u043e\u043b\u0436\u0435\u043d\u0438\u0435\u043c \u0431\u0443\u0434\u0435\u0442 \u0441\u0442\u0430\u0442\u044c\u044f \u043e \u0432\u044b\u044f\u0432\u043b\u0435\u043d\u0438\u0438 \u0430\u043d\u043e\u043c\u0430\u043b\u0438\u0439.","og:url":"https:\/\/prohoster.info\/sv\/blog\/news\/9-podhodov-dlya-vyyavleniya-anomalij","og:image":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:secure_url":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:width":350,"og:image:height":350,"article:published_time":"2019-11-27T21:00:00+00:00","article:modified_time":"2020-02-18T11:01:10+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":{"post_id":"53278","title":null,"description":null,"keywords":null,"keyphrases":null,"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":null,"og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"","isEnabled":true},"graphs":[]},"schema_type":null,"schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":null,"robots_max_videopreview":null,"robots_max_imagepreview":"large","priority":null,"frequency":null,"local_seo":null,"seo_analyzer_scan_date":"2026-01-24 06:46:20","breadcrumb_settings":null,"limit_modified_date":false,"reviewed_by":null,"ai":null,"created":"2021-02-28 20:28:47","updated":"2026-01-24 06:46:20","focus_keyword":null,"additional_keywords":null,"truseo_locale":null},"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/posts\/53278","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/comments?post=53278"}],"version-history":[{"count":0,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/posts\/53278\/revisions"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/media?parent=53278"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/categories?post=53278"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/tags?post=53278"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}