Ururimi AI

BM25 hamwe no gusubiramo amagambo

BM25 nijambo ryibanze ryibanze rishingiye kumurongo urutonde rutanga inyandiko nukuntu inshuro zibazwa zigaragara, zahinduwe kubijyanye nigihe gike nuburebure bwinyandiko.

2 min somaIbiherutse kuvugururwa

Incamake

Decades old, it remains a remarkably strong and ubiquitous baseline for search.

Kwibira cyane

BM25 (Guhuza Byiza 25) nigikapu-yamagambo yo gutondekanya ibikorwa uhereye kuri Okapi ya probabilistic yo mu myaka ya za 90. Kuri buri jambo ryibibazo rihuza ibimenyetso bitatu: inshuro zijambo (inshuro ijambo rigaragara mumyandiko, hamwe no kugabanuka kugaruka kugenzurwa na parameter k1), inshuro zinyuranye zinyandiko (amagambo adakunze kuboneka mubyegeranyo bibarwa cyane), hamwe nuburebure bwinyandiko zisanzwe (parameter b, inyandiko ndende ntabwo zishimirwa). Sum aya manota yigihembwe hanyuma ubone urwego rwinyandiko. Ntabwo ikeneye imyitozo kandi ikora byihuse ikoresheje indangagaciro zahinduwe, niyo mpamvu moteri zishakisha nka Elasticsearch na Lucene zikoresha muburyo budasanzwe. Nubwo kwiyongera kwimitekerereze, BM25 iracyatsinda cyangwa ihuza ibipimo byinshi, cyane cyane kumagambo adasanzwe, ibimuranga neza, hamwe nibibazo bitari hanze.

Ubushishozi

Ikirangantego cya BM25 cyuzuye: ibice bya k1 byerekana uburyo amagambo yasubiwemo azamura amanota, ijambo rero rigaragara inshuro 50 ntabwo ari 50x rifite akamaro kuruta rimwe. B ibipimo bivanga mbisi nuburebure-busanzwe. IDF kumanura-uburemere amagambo asanzwe nka 'the' kandi ihemba ayandi. Kuberako ikora kumurongo uhindagurika ushushanya buri jambo kurutonde rwinyandiko, amanota akora gusa inyandiko zirimo amagambo yibibazo, bigatuma akora neza cyane.

Ingaruka z'Ingamba

Umuvuduko n'igipimo

Ururimi rwakazi rushobora kugenda byihuse nta gutamba guhuzagurika.

Kugera no kugera

Yagura uburyo bwindimi nuburyo bwo gutumanaho.

Ibyemezo bisobanutse

Amakipe arashobora kumara umwanya munini murubanza mugihe automatike ikora gusubiramo.

Kazoza ka BM25 no Kugarura Amagambo

BM25 ntabwo bishoboka ko yazimira; ahubgo biragenda bihuzwa nuburyo bwimitsi muburyo bwo kugarura imvange, aho amanota yamagambo kandi yuzuye ahujwe (akenshi binyuze murwego rwo gusubiranamo). Kwiga gake gake nka SPLADE ivanga imiterere ya BM25 hamwe nuburemere bwimyumvire, kandi BM25 ikora nk'icyiciro cya mbere cyo kugarura mbere yimitsi. Umuvuduko wacyo, gusobanurwa, hamwe namahugurwa ya zeru byemeza uruhare rurambye mugushakisha umusaruro.

Gushyira mu bikorwa Isi

Urutonde rusanzwe rufite akamaro muri Elasticsearch, Gushakisha, na Apache Lucene / Solr

Icyiciro cya mbere cyabakandida kugarura kugaburira buhoro buhoro imitekerereze ya shakisha ibyiciro bibiri

Shakisha kode hamwe nogushakisha aho ibiranga neza na kode yamakosa bigomba guhura neza

Gucukura ingero zikomeye zo guhugura ibyinshi nka DPR

Ingaruka & Kurinda

Ibintu bifatika bishobora kwinjiza bucece raporo, gushyigikira imigendekere, cyangwa ibisubizo byubushakashatsi.

Kwihuta byihuse birashobora gukora ibisubizo bidahuye mubisabwa bisa.

Ibyanditswe byumvikana birashobora kugaragara niba kugenzura kugenzura ari ntege.

Igishushanyo mbonera

1

Sobanura imiterere isohoka, amajwi, hamwe nubuziranenge mbere yo gutangira.

2

Ibisubizo byibanze hamwe nisoko yizewe igihe cyose ukuri kwingirakamaro.

3

Komeza kugenzura abantu kugenzura ibisubizo byinshi.

4

Kurikirana uburyo bwo kunanirwa no kongera imyitozo cyangwa akazi gahoraho.

Komeza Ubushakashatsi

Free newsletter

Get the daily AI briefing

Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.

One email each weekday. Unsubscribe in one click. We never sell or share your address.

Test yourself

Take the BM25 and Lexical Retrieval quiz

Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.

Tangira ikibazo

Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation

Ubuyobozi bukurikira

ColBERT Yatinze Imikoranire

Ibibazo bikunze kubazwa

What is BM25 and Lexical Retrieval?

BM25 nijambo ryibanze ryibanze rishingiye kumurongo urutonde rutanga inyandiko nukuntu inshuro zibazwa zigaragara, zahinduwe kubijyanye nigihe gike nuburebure bwinyandiko. Imyaka mirongo irashize, iracyakomeye bidasanzwe kandi igaragara hose gushakisha.

Niki BM25 ikoresha cyane cyane kurutonde rwinyandiko?

BM25 ikomatanya ijambo inshuro, inverse yinyandiko inshuro (ijambo gake), hamwe nuburebure bwinyandiko mubisanzwe mumanota yingirakamaro.

Ni uruhe ruhare inyandiko zinyuranye (IDF) zigira muri BM25?

IDF ihemba amagambo adakunze gukusanywa hamwe no kumanura-uburemere bw'amagambo asanzwe, kubera ko imikino idasanzwe iratanga amakuru.

Kuki BM25 ikoresha inyandiko-ndende isanzwe (b parameter)?

Hatariho ibisanzwe, inyandiko ndende zegeranya ibihe byinshi bihuye; b ibipimo bihindura uburebure kuburyo kugereranya birakwiye.

Ni ubuhe buryo bw'amakuru butuma BM25 yihuta ku gipimo?

Indangantego ihindagurika ireka BM25 itanga amanota gusa ibyangombwa bikubiyemo amagambo yibibazo, bigatuma kugarura neza.