Nrhiav kom paub tias txoj kev txheeb cais pab kom koj tus neeg tuaj kho kom huv
Bayesian spam lim ntxig xam qhov tshwm sim ntawm cov lus hais tias spam raws li nws cov ntsiab lus. Tsis zoo li cov ntxim nyiam cov ntsiab lus, Bayesian spam filtering kawm tau los ntawm spam thiab los ntawm kev xa ntawv zoo, ua rau muaj kev kub ntxhov, ua kom yooj yim thiab txav txuam rau anti-spam hais tias, qhov zoo tshaj plaws ntawm txhua yam, rov qab tsis nyuaj cov lus dag.
Koj Yuav Ua Li Cas Thiaj Zem Qib Xov Xwm?
Xav txog seb koj ntes tau spam . Ib nrawm nrawm dua yog txaus. Koj paub dab tsi spam zoo li, thiab koj paub tias qhov xa ntawv zoo li cas.
Qhov tshwm sim ntawm spam saib zoo li kev xa ntawv nyob ib ncig ntawm ... xoom.
Txhim Kho Cov Ntsiab Lus Raws Li Txheej Txheem Tsis Tau
Nws yuav tsis zoo yog hais tias siv neeg tua phom nyiag tuaj ua haujlwm zoo li ntawd, thiab?
Txheeb zus cov ntsiab lus raws li kev xaiv spam sim li ntawd. Lawv nrhiav cov lus thiab lwm yam ntxwv uas yog spam. Txhua tus yam ntxwv ntawm tus cwj pwm tau muab rau tus qhab nia, thiab ib qho txiaj ntsim ntawm qhov spam rau tag nrho cov lus yog xam los ntawm cov qhab nia ntawm tus kheej. Qee cov cim ntxaiv kuj nrhiav tau cov yam ntxwv ntawm cov ntawv xa ncaj qha, txo nqe lus kawg nkaus.
Cov kev ntxig ntxig ntxig ntxig rau kev ua haujlwm ua haujlwm, tabsis nws kuj muaj ntau qhov teebmeem:
- Cov npe ntawm cov cwj pwm yog tsim los ntawm spam (thiab xa ntawv zoo) rau cov lim cov engineers. Yuav kom paub tseeb txog qhov raug spam tus neeg muaj peev xwm tau txais, xa ntawv yuav tsum tau sau rau ntau pua email chaw nyob. Qhov no ua rau cov neeg lim dej ua haujlwm tsis zoo, tshwj xeeb tshaj yog vim tias cov kev xa ntawv zoo yuav txawv rau txhua tus neeg , tab sis qhov no tsis raug coj mus rau hauv.
- Cov yam ntxwv zoo rau kev nrhiav tau ntau lossis tsawg dua teeb tsa hauv pob zeb . Yog tias cov spammers siv zog los hloov (thiab ua rau lawv cov spam zoo li xa ntawv xa mus rau cov ntxaij lim dej), tus cwj pwm filtering yuav tsum tau tweaked manually - ua haujlwm ntau dua.
- Tus qhab nia muab rau txhua lo lus yog zaum ntawm ib qho kev ntsuas zoo, tiam sis nws tseem yog qhov tseem ceeb. Thiab zoo li daim ntawv teev cov yam ntxwv, nws tsis hloov lossis hloov lub ntiaj teb ntawm spam feem ntau los yog rau tus neeg siv cov kev xav tau.
Bayesian Spam Lim Tweak Themselves, Tau Zoo Tshaj thiab Zoo
Bayesian spam cov ntxaij lim dej yog ib hom kev teev ntsiab lus ntawm cov yeeb yaj duab, thiab. Lawv mus kom ze nrog cov teeb meem ntawm cov lus cim yooj yim spam, kiag, thiab nws ua li ntawd radically. Txij li cov tsis muaj zog ntawm cov ntxaij lim dej yog nyob rau hauv daim ntawv teev kev ua manually ua cov yeeb yam thiab lawv cov qhab nias, cov npe no tshem tawm.
Es tsis txhob, Bayesian spam lim tsim cov npe lawv tus kheej. Hom phiaj, koj pib nrog ib pawg (loj) ntawm cov emails uas koj tau muab cais ua spam, thiab lwm pawg xa ntawv zoo. Cov ntxaij lim dej tseem pom ob qho thiab tshawb xyuas cov ntawv pov thawj raws li lub spam los xam qhov tshwm sim ntawm ntau yam yeeb yam uas tshwm hauv spam, thiab xa ntawv zoo.
Yuav ua li cas ib Bayesian Spam Lim kuaj ib tug Email
Cov yam ntxwv ntawm Bayesian spam lim yuav saib tau:
- cov lus nyob rau hauv lub cev ntawm cov lus, ntawm chav kawm, thiab
- nws cov headers (xa thiab lus paths , piv txwv li!), tab sis kuj
- lwm yam xws li HTML / CSS code (xws li cov xim thiab lwm yam formatting), los sis txawm
- lo lus khub, nqe lus thiab
- meta cov ntaub ntawv (qhov twg ib qho kab lus tshwm, piv txwv).
Yog hais tias lo lus, "Cartesian" piv txwv li, yeej tsis tshwm rau hauv spam, tiam sis feem ntau nyob rau hauv qhov tseeb email koj tau txais, qhov tshwm sim hais tias "Cartesian" qhia tias spam yog nyob ze ntawm zero. "Toner", ntawm qhov tod tes, zoo li ntau dua, thiab feem ntau, hauv spam. "Toner" muaj ib qho siab heev ntawm kev pom nyob rau hauv spam, tsis ntau tshaj 1 (100%).
Thaum ib cov lus tshiab tuaj txog, nws tau soj ntsuam los ntawm Bayesian spam lim, thiab qhov tshwm sim ntawm tag nrho cov lus ua spam yog xam siv tus neeg cov cwj pwm.
Cia siab tias cov lus muaj ob lo lus "Cartesian" thiab "toner". Los ntawm cov lus no ib leeg tseem tsis tau paub tseeb tias peb muaj cov ntawv xa email los yog cov ntawv xa tawm. Lwm cov cwj pwm yuav ua tau (hopefully thiab feem ntau tej zaum) qhia tias yog qhov muaj tseeb uas tso cai rau lim kom faib tawm lus raws li spam los yog xa ntawv zoo.
Bayesian Spam Lim tau Kawm Tau
Tam sim no hais tias peb muaj ib qho kev faib tawm, cov lus siv tau los cob qhia cov lim nws tus kheej ntxiv. Hauv qhov no, yog qhov tshwm sim ntawm "Cartesian" qhia tias zoo xa ntawv yog txo (yog hais tias cov lus uas muaj "Cartesian" thiab "toner" yog pom tias spam), los yog qhov tshwm sim ntawm "toner" qhia tias spam yuav tsum rov qab txheeb xyuas.
Siv cov txheej txheem nws pib ua haujlwm, Bayesian cov ntxoov ntxoo tuaj yeem kawm tau los ntawm lawv tus kheej thiab tus neeg siv qhov kev txiav txim siab (yog tias nws tswj kev txiav txim siab ntawm cov ntxaij lim dej). Tus adaptability ntawm Bayesian filtering kuj ua rau kom lawv zoo tshaj rau tus neeg siv email. Txawm hais tias feem coob tus neeg spam tej zaum yuav muaj cov yam ntxwv zoo sib xws, qhov tseeb ntawv yog txawv dua rau txhua tus.
Npaum Cas Tau Spammers Tau Txais Bayesian Lim Yav dhau los?
Cov cwj pwm ntawm cov ntawv xa ncaj qha yog ib qho tseem ceeb rau txoj kev ntaus cim Bayesian spam raws li cov ntawv tseem ceeb. Yog hais tias cov ntxaij lim dej raug cob qhia rau txhua tus neeg, spammers yuav muaj kev nyuaj rau lub sijhawm ua haujlwm rau txhua tus neeg (lossis cov neeg feem coob) spam cov ntxaij lim dej, thiab cov ntxaij lim dej tuaj yeem hloov tau txhua yam lus nug.
Spammers yuav tsuas yog ua kom dhau keeb kwm zoo nkauj kawm txog Bayesian yog tias lawv ua lawv cov lus spam zoo zoo nkaus li qhov zoo tib yam email txhua tus neeg tau txais.
Spammers tsis xa cov npe zoo li no. Cia peb xav tias qhov no yog vim tias cov emails tsis ua haujlwm li khoob email. Yog li, muaj feem ntau lawv yuav tsis ua nws thaum zoo tib yam, tho txawv emails yog tib txoj kev los ua nws dhau spam lim.
Yog hais tias spammers hloov mus rau cov saibxyuas dog dig zoo saib, txawm li cas los, peb yuav pom ntau spam nyob rau hauv peb Inboxes dua, thiab email tej zaum yuav ua rau kev ntxhov siab thaum nws nyob rau hauv pre-Bayesian hnub (los yog txawm tshaj ntawd). Nws tseem yuav muaj kev puas tsuaj rau kev ua lag luam rau ntau hom kev spam, tab sis yog li ntawd tsis ntev li ntev.
Cov Poj Cov Kev Ntsuas Yuav Ua Tau Ib Bayesian Spam Lim & # 39; Achilles & # 39; Luj
Ib qho kev pom zoo yuav raug ntaus nqi rau cov spammers los ua haujlwm rau lawv txoj kev los ntawm Bayesian cov ntxaij lim dej nrog lawv cov txheej txheem li ib txwm. Nws yog nyob rau ntawm Bayesian statistics uas ib lo lus los yog yam ntxwv uas nquag pom hauv kev xa ntawv tuaj yeem tseem ceeb ua lem lus los ntawm kev saib zoo li spam mus ua qhwv li ham los ntawm lub lim.
Yog tias spammers nrhiav txoj hau kev txiav txim siab rau koj cov lus pov thawj zoo-xa ntawv-siv cov ntawv xa nyiaj los ntawm HTML los xa cov lus uas koj tau qhib, piv txwv li, lawv tuaj yeem muaj xws li ib qho ntawm cov ntawv xa mus khoob khaus thiab tuaj yeem hu koj los ntawm kev ua tau zoo- kawm tiav Bayesian.
John Graham-Cumming tau sim ua qhov no los ntawm ob daim Bayesian cov duab ua haujlwm sib tawm tsam, qhov "phem" yog ib qho kev sib txuas lus uas cov lus tau pom los ntawm "lim" zoo. Nws hais tias nws ua haujlwm, tab sis yog qhov txheej txheem yog siv sij hawm thiab tsis yooj yim. Peb tsis xav tias peb yuav pom ntau qhov teeb meem no, tsawg kawg tsis yog nyob rau hauv lub loj teev, thiab tsis yog rau cov tib neeg tus qauv hauv email. Spammers yuav (sim) xam ib co ntsiab lus tseem ceeb rau cov koom haum (tej yam xws li "Almaden" rau qee cov neeg ntawm IBM tej zaum?) Xwb.
Feem ntau, spam yuav zoo ib yam (yawg) txawv ntawm kev xa ntawv lossis nws yuav tsis yog spam, ho.
Tus Kab Hauv qab: Bayesian Filtering & # 39; Strength Ua Tau Nws Tsis Ncaj Ncees
Bayesian spam cov ntxaij lim dej yog cov ntsiab lus raws li cov lim uas:
- yog cov tau txais kev cob qhia kom paub txog tus kheej email tus neeg siv lub spam thiab xa ntawv zoo , ua rau lawv siv tau zoo thiab nyuab yoog mus rau spammers.
- tuaj yeem tsis tu ncua thiab tsis muaj kev sib zog los yog kev tsom xam cov lus siv hloov mus rau cov lus cog lus tshiab.
- coj tus neeg siv tus xa ntawv zoo xa nyiaj mus rau hauv tus account thiab muaj feem tsawg heev ntawm cov lus cuav positives .
- Hmoov tsis, yog tias qhov no ua rau cov neeg dig muag pom kev ntseeg hauv Bayesian anti-spam lim, nws ua qhov yuam kev li niaj zaus . Cov nyhuv txawv ntawm cov lus tsis muaj tseeb (spam uas zoo ib yam li cov ntawv xa tuaj) muaj peev xwm ua kom muaj kev ntxhov siab thiab ntxhov siab rau cov neeg siv.