Txheej txheem cej luam: Nrog ces kev coj ib ncig, Nws tseem ceeb heev uas yuav tau teev rau saum peb cov ntaub ntawv ntawm tej taw tes muab sij hawm yuav. Hom kev kawm kawm tshuab clustering thiab categorization kuj nrov no lub ntsiab lus teb. Apache Mahout yog ib lub moj khaum uas pab peb after scalability.
Nyob rau hauv daim ntawv no, Kuv yuav tham txog Apache Mahout thiab qhov tseem ceeb.
Taw qhia: Apache Mahout yog ib qhov qhib tau qhov twg los Apache Software Foundation los yog muaj tus thawj lub hom phiaj ntawm txoj kev siv lub tshuab kev kawm algorithm kom ASF. Ua tswvcuab yog ib pawg developers ntawm peb tes num Apache Lucene, Apache Mahout muaj tus taw rau –
- Tsim thiab txhawb lub zej zog cov neeg los yog contributors kom thiaj mus saib tau hauv qhov chaws rau lub moj khaum tsi muaj cwj ciam rau ib pab pawg me ntawm developers.
- Siab rau cov teeb meem uas ua tau, qhov teeb meem unseen los yog unproved.
- Muab lwm cov ntaub ntawv tsim nyog.
Nta cov Apache Mahout:
Apache Mahout los nrog ib tug array nta thiab functionalities tshwj xeeb tshaj yog thaum peb tham txog Clustering thiab cob qha Filtering. Cov yam ntxwv tseem ceeb tshaj plaws yog li teev kom muaj –
- Saj yaam Filtering – Saj yog ib qhov qhib tau qhov twg los kev cob qha filtering. Nws yog ib sab ntawm lub moj Mahout khaum uas siv tshuab kawm algorithms mus teev txog peb kev siv. Saj yog siv rau koj tus kheej seb. Hnub no thaum peb qhib ib lub website peb paub kom ntau tswv yim hais txog lub website uas peb browsing. Hauv daim duab hauv qab no qhia tau hais tias daim duab architecture ntawm saj –
Daim duab Architecture saj
Daim duab 1: Daim duab Architecture saj
- Daim ntawv qhia kom enabled implementations – Txo enabled ob peb daim ntawv qhia clustered implementations txaus siab nyob rau hauv Mahout. Qhov no K-txhais, fuzzy, Toj
- Faib Navie Bayes thiab Complimentary Navie Bayes- Apache mahout tau qhov kev siv rau Navie Bayes thiab Complimentary Bayes. Rau simplicity Navie bayes yuav xa koj mus li Bayes thiab Complimentary yog hu ua CBayes. Bayes siv nyob rau hauv phau ntawv kev faib tawm thaum lub CBayes extension cov Bayes uas yog siv thaum 'Datasets'.
- Nws txhawb Matrix thiab lwm lwm yam vector qiv.
Qhov pem Apache Mahout:
Teev tau Apache Mahout yooj yim heev thiab yuav muab nqa ntawm cov kauj ruam hauv qab no –
- Kauj ruam 1 – Rau order los teeb Apache Mahout, peb yuav tsum tau muaj hauv qab no ntsia –
- JDK 1.6 lossis siab dua ntawd
- Ntsaum 1.7 lossis siab dua ntawd
- Maven 2.9 lossis siab dua ntawd – thaum peb xav tsim tawm hauv qhov chaws
- Kauj ruam 2 -Unzip ntawv, Sample.Zip thiab daim ntawv tus txheem hauv tej sis hais "apache-mahout-piv txwv".
- Kauj ruam 3 – Nkag mus rau hauv daim nplaub tshev – "apache-mahout-piv txwv" thiab cov nram no mus dhia –
- ntsaum nruab
Cov kauj ruam dhau downloads cov ntaub ntawv Wikipedia thiab compiles tus.
Xav cav:
Xav cav yog ib subclass filtering uas twv seb cov ntsuam xyuas tau lawv cov lus los yog nyiam cov neeg siv yuav muab rau ib yam khoom. Mahout muaj cuab yeej thiab tej hom kev kawm uas yuav pab tau kom lawv muaj kev pom zoo xyaw siv lub tsev qiv ntawv 'Saj'. Siv tau saj tsev qiv ntawv peb yuav tsa ib tug ceev thiab saj zawg zog yaam Filtering cav. Saj muaj cov nram qab no tsib hom cheebtsam uas ua hauj lwm nrog cov neeg, yam khoom thiab nyiam –
- Hom ntaub ntawv – Qhov no yuav muab siv ua rau lub cia lawv cov neeg, yam khoom thiab kuj nyiam.
- Neeg siv kev zoo sib thooj – Qhov no yog ib qho interface uas siv kom txhais tau cov kev zoo sib thooj ntawm cov neeg ob.
- Yam kev zoo sib thooj – Tus interface uas siv kom txhais tau cov kev zoo sib thooj ntawm ob yam khoom.
- Recommender – Tus interface uas yog siv los muab tswv yim.
- Neeg zej zog neeg siv – Tus interface uas yog siv los laij thiab laij ib cov neeg zej zog rau cov neeg siv ntawm qeb ib yam uas yuav siv tau los ntawm cov Recommenders.
Siv tej lub Cheebtsam thiab lawv cov implementations, peb yuav tsim ib txoj kev pom zoo lawv. No xav cav yuav siv tau rau cov sij hawm ntawm tswv yim thiab tswv yim offline. Lub sij hawm ntawm tswv yim yuav lis cov neeg mus txog ob peb txhiab thaum offline pom zoo yuav lis cov neeg nyob hauv suav ntau npaum li cas.
Clustering:
Mahout txhawb ntau clustering mechanisms. Cov algorithms sau rau hauv daim ntawv qhia kom tsis txhob. Txhua yam ntawm cov algorithms muaj nyias txheej cov hom phiaj thiab kev. Qhov tseem ceeb sawv daws yuav muaj teev nyob kom muaj-
- Toj- Qhov no yog tus ceev tshaj clustering algorithm siv los ua thawj zaug muaj noob rau lwm algorithms clustering.
- k – Txhais tau tias yog Fuzzy k – txhais hais tias- Cov algorithm no ua ib qho k nyob raws li cov kev coj hauv cov khoom ntawm tus centre ntawm tus iteration yav dhau los.
- Txhais hais tias-cov ua haujlwm – Cov algorithm no tsis tau muaj ntaub ntawv ua ntej txog pes tsawg tus nyob. Qhov no yuav ua ib pawg arbitrary uas yuav muab ntau zog los yog tsis as per peb yuav tau.
- Dirichlet – Cov algorithm no ua ib qho kev nyob ntawm combining ib los sis ntau dua pawg qauv. Yog li peb tau txais ib qho kom zoo dua rau xaiv zoo tau ib tug los ntawm tej pawg ua ke ntau.
Tawm hauv plaub lub algorithms saum toj no nyob, qhov ntau yog lub k – txhais tau tias algorithm. Muab nws tej clustering algorithm, peb yuav tsum ua raws li tej kauj ruam nram –
- Npaj lub tswv yim. Yog hais tias yuav tsum tau, hloov ua lwm yam hauv cov ntawv nyeem rau sawv cev numeric.
- Coj tus algorithm uas koj xaiv los ntawm qhov uas siv yam lub Hadoop npaj cov kev pab cuam uas muaj nyob rau hauv Mahout.
- Kom luj xyuas cov ntsiab.
- Iterate cov kauj ruam nram no yog hais tias yuav tsum tau.
Cov Categorizing:
Apache Mahout txhawb cov nram qab no ob le caag categorize los yog cov lus raws li lub. Cov no mas raws li txheeb cais Bayesian-
- Cov thawj kev no daim ntawv qhia ncaj pem txo rau enabled Navie bayes classifier. Classifiers ntawm pawg no paub tias yuav tsum ceev thiab meej dua li muaj tus assumption tias cov ntaub ntawv yog neeg sab nraud kiag li. Cov classifiers zom thaum ntawd luaj li cas ntawm cov ntaub ntawv mus los yog cov ntaub ntawv yuav interdependent. Navie bayes classifier yog muaj ob theem uas yuav muaj ib tug khiav ntawm cov yam ntxwv los yog ib tsab cov lus uas thaum uas ib qho ntaub ntawv. Cov kauj ruam no yog paub tias thaum uas tseem ua ib qho tus qauv los saib cov piv txwv uas twb kawm ntawv txwv kom muab zais. Cov kauj ruam thib ob, peb paub los lawm raws li kev faib tawm, siv tus qauv uas yog tsim nyob rau txoj kev kawm thiab cov ntsiab lus ntawm ib tug tshiab, unseen tsab ntawv. Li no, thiaj li khiav lub Mahout classifier, Peb xub tau qhia tus qauv thiab ces siv tus qauv kom muaj qeb muaj quag tshab.
- Qhov thib ob mus kom ze, Nws yog ib qho los yog Complementary Naive Bayes, yuav muaj qee yam uas hais txog cov lus uas cov Naive Bayes kev rectify thiab tseem koom tes hauv simplicity thiab ceev saib xyuas Navie Bayes.
Si laim cov Navie Bayes Classifier:
Cov Navie Bayes Classifier kom executing hauv qab no lub hom phaj ntsaum thiaj li coj –
- ntsaum npaj-docs – Npaj tej ntaub ntawv uas yuav tsum tau kawm lub txheej qhov no.
- ntsaum npaj-no-docs – Npaj cov ntaub ntawv uas yuav tsum tau mus kuaj txheej qhov no.
- tsheb nqaj hlau ntsaum -Thaum cov kev kawm thiab cov kev ntsuam xyuas cov ntaub ntawv teev, peb yuav khiav cov hoob uas siv cov phiaj – "ntsaum hlau" TrainClassifier.
- ntsaum kuaj – Thaum tus saum lub hom phaj muaj ntawv kho mob tseg ntse, peb yuav tsum khiav no phiaj uas yuav siv tus qauv input ntaub ntawv thiab yuav faib lawv ua raws li cov qauv uas yog tsim thaum uas cob qhia.
Txoj kev: Nyob rau cov tshooj no peb pom tias Apache Mahout lug siv rau phau ntawv kev faib tawm thaum uas siv cov tshuab kawm algorithms. Lub tshuab tseem loj hlob thiab siv tau rau daim ntawv thov kev sib txawv. Peb cia saib lub ntsab lug peb sib tham nyob rau hauv daim ntawv uas cia nyias qhov hauv qab no –
- Apache Mahout yog ib qhov qhib tau qhov twg los txhob nkag tau los ntawm ib pawg developers ntawm peb tes num Apache Lucene Apache. Thawj lub hom phiaj rau qhov project no yog tsim algorithm uas tau nyeem lus cav.
- Apache Mahout muaj tus tseem ceeb nta hauv qab no –
- Saj yaam Filtering.
- MapReduce enabled implementations.
- Yuav siv rau muab theej thiab faib Navie Bayes thiab Complimentary Navie Bayes.
- Matrix txhawb zog thiab lwm yam raws li vector qiv.
