Okuqukethwe[Fihla][Bonisa]
Phela, sonke siyazi ukuthi ubuchwepheshe bokufunda ngomshini buthuthuke ngokushesha kangakanani eminyakeni embalwa edlule. Ukufunda ngomshini kuyisiyalo esihehe intshisekelo yezinkampani ezimbalwa, izifundiswa, nemikhakha.
Ngenxa yalokhu, ngizoxoxa ngezinye zezincwadi ezinhle kakhulu zokufunda ngomshini okufanele zifundwe unjiniyela noma osanda kuzalwa namuhla. Kumele ngabe nivumile nonke ukuthi ukufunda izincwadi akufani nokusebenzisa ubuhlakani.
Ukufunda izincwadi kusiza izingqondo zethu ukuthola izinto eziningi ezintsha. Ukufunda kuwukufunda phela. Ithegi yokuzifundela kumnandi kakhulu ukuba nayo. Izincwadi zokufunda ezinkulu kunazo zonke ezitholakala ensimini zizoqokonyiswa kulesi sihloko.
Lezi zincwadi ezilandelayo zinikeza isingeniso esizanyiwe neseqiniso emkhakheni omkhulu we-AI futhi zivame ukusetshenziswa ezifundweni zasenyuvesi futhi zituswe izifundiswa nonjiniyela ngokufanayo.
Ngisho noma unethani ukufunda imishini isipiliyoni, ukucosha enye yalezi zincwadi kungase kube indlela enhle kakhulu yokuxubha. Phela ukufunda kuyinqubo eqhubekayo.
1. Ukufunda Ngomshini Kwabaqalayo Abaphelele
Ungathanda ukufunda ukufunda ngomshini kodwa awazi ukuthi ukwenze kanjani. Kunemiqondo eminingana ebalulekile yethiyori neyezibalo okufanele uyiqonde ngaphambi kokuqala uhambo lwakho oludumile lokufunda ngomshini. Futhi le ncwadi iyasigcwalisa leso sidingo!
Inikeza ama-novices aphelele anezinga eliphezulu, elisebenzayo isingeniso sokufunda ngomshini. Incwadi ethi Machine Learning for Absolute Beginners ingenye yezinketho ezingcono kakhulu kunoma ubani ofuna incazelo eyenziwe lula yokufunda ngomshini nemibono ehlobene.
Ama-algorithms amaningi wencwadi ahambisana nezincazelo ezimfushane nezibonelo eziyingcaca ukusiza abafundi baqonde yonke into okuxoxwa ngayo.
Izihloko ezifakwe encwadini
- Izisekelo ze amanethiwekhi we-neural
- Ukuhlaziywa komthetho
- Ubunjiniyela besici
- Ukuhlanganisa
- Ukuqinisekisa okuphambene
- Amasu okukhuhla idatha
- Izihlahla Zesinqumo
- Amamodeli ahlanganisiwe
2. Ukufunda ngomshini kwama-Dummies
Ukufunda ngomshini kungase kube umqondo odidayo kubantu abavamile. Nokho, iyigugu kithina esinolwazi.
Ngaphandle kwe-ML, kunzima ukuphatha izindaba ezifana nemiphumela yosesho lwe-inthanethi, izikhangiso zesikhathi sangempela emakhasini ewebhu, i-automation, noma ukuhlunga ogaxekile (Yebo!).
Njengomphumela walokho, leli bhuku likunikeza isingeniso esiqondile esizokusiza ufunde kabanzi mayelana nendawo eyindida yokufunda komshini. Ngosizo lwe- Machine Learning For Dummies, uzofunda ukuthi "ungakhuluma" kanjani izilimi ezifana ne-Python ne-R, okuzokwenza ukwazi ukuqeqesha amakhompyutha ukwenza ukuqashelwa kwephethini nokuhlaziya idatha.
Ukwengeza, uzofunda ukusebenzisa i-Python's Anaconda ne-R Studio ukuthuthukisa ku-R.
Izihloko ezifakwe encwadini
- Ukulungiswa kwedatha
- izindlela zokufunda ngomshini
- Umjikelezo wokufunda komshini
- Ukufunda okugadiwe nokungagadiwe
- Amasistimu okufunda emishini yokuqeqesha
- Ukubophela izindlela zokufunda zomshini emiphumeleni
3. Incwadi Yokufunda Ngomshini Amakhasi Ayikhulu
Ingabe kuyenzeka ukumboza zonke izici zokufunda komshini ngaphansi kwamakhasi ayi-100? Incwadi ka-Andriy Burkov ethi The Hundred-page Machine Learning Book ingumzamo wokwenza okufanayo.
Incwadi yokufunda ngomshini ibhalwe kahle futhi isekelwa abaholi bemicabango abadumile abahlanganisa u-Sujeet Varakhedi, iNhloko Yobunjiniyela e-eBay, no-Peter Norvig, uMqondisi Wocwaningo kwa-Google.
Liyincwadi enkulu kunazo zonke yomuntu oqalayo ekufundeni ngomshini. Ngemuva kokuyifunda kahle le ncwadi, uzokwazi ukwakha nokuqonda amasistimu e-AI asezingeni eliphezulu, uphumelele kwinhlolokhono yokufunda ngomshini, futhi uze wethule eyakho inkampani esekwe ku-ML.
Kodwa-ke, le ncwadi ayihloselwe abaqalayo abaphelele ekufundeni komshini. Bheka kwenye indawo uma ufuna okuthile okubaluleke kakhulu.
Izihloko ezifakwe encwadini
- I-anatomy ye- a i-algorithm yokufunda
- Ukufunda okugadiwe nokufunda okungagadiwe
- Ukuqinisa Ukufunda
- Ama-algorithms ayisisekelo wokufunda ngomshini
- Ukubuka konke kwamanethiwekhi e-Neural nokufunda okujulile
4. Ukuqonda Ukufunda Ngomshini
Isingeniso esihlelekile sokufunda ngomshini sinikezwa encwadini ethi Ukuqonda Ukufunda Ngomshini. Le ncwadi icubungula ngokujulile emibonweni eyisisekelo, ama-paradigms ekhompyutha, nokuphuma kwezibalo zokufunda komshini.
Uhlu olubanzi lwezifundo zokufunda ngomshini lwethulwa ngendlela elula ngokufunda komshini. Izisekelo zetiyori zokufunda komshini zichazwe encwadini, kanye nokuphuma kwezibalo okushintsha lezi zisekelo zibe ama-algorithms awusizo.
Le ncwadi yethula okuyisisekelo ngaphambi kokuhlanganisa izihloko eziningi ezibalulekile ezingakaze zenziwe ezincwadini zangaphambili.
Kufakwe kulokhu ingxoxo ye-convexity kanye nokuqina kwemiqondo kanye nobunkimbinkimbi bekhompyutha bokufunda, kanye nama-algorithmic paradigms abalulekile afana ne-stochastic. ukwehla kwe-gradient, amanethiwekhi we-neural, nokufunda okuphumayo okuhlelekile, kanye nemibono esanda kuvela yasetiyetha efana nendlela ye-PAC-Bayes kanye nemingcele esekelwe ekucindezelweni. yakhelwe ama-grade aqalayo noma abenza iziqu ezithuthukile.
Izihloko ezifakwe encwadini
- Ubunkimbinkimbi bekhompyutha bokufunda komshini
- ML algorithms
- Ama-Neural amanethiwekhi
- Indlela ye-PAC-Bayes
- Ukwehla kwe-Stochastic gradient
- Ukufunda kokuphumayo okuhlelekile
5. Isingeniso Sokufunda Ngomshini ngePython
Ingabe ungusosayensi wedatha wePython-savvy ofuna ukufunda ngomshini? Ibhuku elingcono kakhulu ongaqalisa ngalo uhambo lwakho lokufunda ngomshini Isethulo Sokufunda Ngomshini ngePython: Umhlahlandlela Wososayensi Bedatha.
Ngosizo lwencwadi ethi Isingeniso Sokufunda Ngomshini NgePython: Umhlahlandlela Wososayensi Bedatha, uzothola izindlela ezihlukahlukene eziwusizo zokudala izinhlelo zokufunda zomshini ngokwezifiso.
Uzofaka zonke izinyathelo ezibalulekile ezihilelekile ekusebenziseni i-Python kanye nephakheji ye-Scikit-Learn ukuze wakhe izinhlelo zokusebenza zokufunda zomshini ezithembekile.
Ukuthola ukuqonda okuqinile kwemitapo yolwazi ye-matplotlib ne-NumPy kuzokwenza ukufunda kube lula kakhulu.
Izihloko ezifakwe encwadini
- Izindlela zesimanje zokulungisa ipharamitha nokuhlola imodeli
- Izicelo kanye nemibono eyisisekelo yokufunda komshini
- amasu okufunda okuzenzakalelayo
- Amasu okukhohlisa idatha yombhalo
- Amapayipi we-chaining kanye nokuhamba komsebenzi we-encapsulation
- Ukumelwa kwedatha ngemva kokucubungula
6. Ukufunda ngomshini ngezandla nge-Sci-kit yokufunda, i-Keras ne-Tensorflow
Phakathi kokushicilelwe okuphelele kwesayensi yedatha nokufunda komshini, igcwele ulwazi. Kuyalulekwa ukuthi ochwepheshe kanye nabaqalayo ngokufanayo bafunde kabanzi ngalesi sihloko.
Nakuba le ncwadi iqukethe ithiyori encane nje, isekelwa izibonelo ezinamandla, ezinikeza indawo ohlwini.
Leli bhuku lihlanganisa izihloko ezihlukahlukene, ezihlanganisa i-scikit-lear yamaphrojekthi okufunda ngomshini kanye ne-TensorFlow yokudala nokuqeqesha amanethiwekhi e-neural.
Ngemva kokufunda le ncwadi, sicabanga ukuthi uzohlonyiswa kangcono ukuze uhlole kabanzi ukufunda okujulile futhi ubhekane nezinkinga ezingokoqobo.
Izihloko ezifakwe encwadini
- Hlola isimo sokufunda komshini, ikakhulukazi amanethiwekhi e-neural
- Landelela isampula yephrojekthi yokufunda komshini kusukela ekuqaleni kuya esiphethweni usebenzisa i-Scikit-Learn.
- Hlola amamodeli wokuqeqesha amaningana, njengamaqhinga wokuhlanganisa, amahlathi angahleliwe, izihlahla zezinqumo, nemishini yokusekela ye-vector.
- Dala futhi uqeqeshe amanethiwekhi e-neural ngokusebenzisa umtapo wezincwadi we-TensorFlow.
- Cabangela amanethiwekhi okuxhumana, amanethi aphindelelayo, nokufunda okujulile kokuqinisa ngenkathi uhlola inetha ye-neural imiklamo.
- Funda ukuthi ungakala kanjani futhi uqeqeshe amanethiwekhi e-neural ajulile.
7. Ukufunda ngomshini kwabaduni
Ngomhleli onolwazi onentshisekelo yokuhlaziya idatha, incwadi ethi Machine Learning for Hackers ibhaliwe. Abaduni bezibalo bangochwepheshe bezibalo kulo mongo.
Kumuntu onokuqonda okuqinile kwe-R, leli bhuku liyinketho enhle ngoba iningi lalo ligxile ekuhlaziyweni kwedatha ku-R. Ukwengeza okufakwe encwadini ukuthi ungakhohlisa kanjani idatha usebenzisa i-Advanced R.
Ukufakwa kwezindaba zamacala abalulekile kugcizelela ukubaluleka kokusebenzisa ama-algorithms okufunda komshini kungaba ibhuku elithi Machine Learning for Hackers' indawo ebaluleke kakhulu yokuthengisa.
Leli bhuku linikeza izibonelo eziningi zomhlaba wangempela ukwenza ukufunda ngomshini wokufunda kube lula futhi kusheshe kunokungena ujule kuthiyori yayo yezibalo.
Izihloko ezifakwe encwadini
- Dala isigaba se-Bayesian esingenangqondo esihlaziya kalula okuqukethwe kwe-imeyili ukuze unqume ukuthi ugaxekile yini.
- Ukubikezela inani lokubukwa kwekhasi kumawebhusayithi aphezulu ayi-1,000 kusetshenziswa ukuhlehla komugqa
- Phenya izindlela zokuthuthukisa ngokuzama ukuhlukanisa uhlamvu lwe-cipher oluqondile.
8. Python Machine Learning with Izibonelo
Le ncwadi, ekusiza ukuthi uqonde futhi udale izindlela ezahlukahlukene Zokufunda Ngomshini, Ukufunda Okujulile, kanye nezindlela Zokuhlaziya Idatha, cishe iyona kuphela egxile kuPython kuphela njengolimi lokuhlela.
Ihlanganisa amalabhulali amaningana anamandla okusebenzisa ama-algorithms ahlukene wokufunda ngomshini, njenge-Scikit-Learn. Imojula ye-Tensor Flow ibe isisetshenziswa ukukufundisa ngokufunda okujulile.
Okokugcina, ibonisa amathuba amaningi okuhlaziya idatha angazuzwa kusetshenziswa umshini nokufunda okujulile.
Iphinde ikufundise izindlela eziningi ezingasetshenziswa ukukhulisa ukusebenza kahle kwemodeli oyidalayo.
Izihloko ezifakwe encwadini
- I-Python Yokufunda Nokufunda Ngomshini: Umhlahlandlela Wabaqalayo
- Ihlola isethi yedatha yamaqembu ezindaba angu-2 kanye nokutholwa kwe-imeyili yogaxekile ye-Naive Bayes
- Usebenzisa ama-SVM, hlela izihloko zezindaba zezindaba ngokuchofoza-ngokuqagela usebenzisa ama-algorithms asekelwe ezihlahleni
- Ukubikezela kwezinga lokuchofoza kusetshenziswa ukuhlehla kwezinto
- Ukusetshenziswa kwe-algorithms yokwehla ukuze kubikezelwe amazinga aphezulu kakhulu wezintengo zesitoko
9. Python Machine Learning
Incwadi yePython Machine Learning ichaza izinto eziyisisekelo zokufunda ngomshini kanye nokubaluleka kwayo esizindeni sedijithali. Liyincwadi yokufunda ngomshini yabaqalayo.
Ukwengeza okufakwe ebhukwini ukufunda ngomshini izinkundla eziningi ezingaphansi nezinhlelo zokusebenza. Izimiso zohlelo lwePython nokuthi ungaqala kanjani ngolimi lokuhlela lwamahhala nomthombo ovulekile nazo zifakiwe encwadini yokufunda yomshini wePython.
Ngemva kokuqeda incwadi yokufunda yomshini, uzokwazi ukusungula ngempumelelo inombolo yemisebenzi yokufunda ngomshini usebenzisa ikhodi yePython.
Izihloko ezifakwe encwadini
- I-Artificial intelligence fundamentals
- isihlahla sesinqumo
- Ukuhlehla kwezinto
- Amanethiwekhi we-neural ajulile
- Izisekelo zolimi lohlelo lwePython
10. Ukufunda Ngomshini: Umbono Onokwenzeka
Ukufunda Ngomshini: Umbono Ongangenzeka yincwadi yokufunda yomshini ehlekisayo enezithombe ezinemibala engavamile kanye nezibonelo ezingokoqobo, zomhlaba wangempela ezivela emikhakheni efana nebhayoloji, umbono wekhompyutha, amarobhothi, nokucubungula umbhalo.
Igcwele i-casual prose kanye ne-pseudocode yama-algorithms abalulekile. Ukufunda Ngomshini: Umbono Onokwenzeka, ngokungafani nokunye okushicilelwe komshini okwethulwa ngesitayela sencwadi yokupheka futhi kuchaze izindlela ezihlukile ze-heuristic, kugxile endleleni esekelwe kumodeli.
Icacisa amamodeli we-ml kusetshenziswa izethulo ezinemifanekiso ngendlela ecacile neqondakalayo. Ngokusekelwe endleleni ehlangene, enokwenzeka, le ncwadi inikeza isingeniso esiphelele nesizimele endaweni yokufunda komshini.
Okuqukethwe kubanzi futhi kujulile, okuhlanganisa okubalulekile okungemuva ngezihloko ezifana namathuba, ukwenziwa ngcono, kanye ne-algebra yomugqa, kanye nengxoxo yentuthuko yesimanje endaweni njengezinkambu ezingahleliwe ezinemibandela, ukujwayela kwe-L1, nokufunda okujulile.
Incwadi ibhalwe ngolimi olujwayelekile, olungenekayo, oluqukethe i-pseudo-code yama-algorithms abalulekile abalulekile.
Izihloko ezifakwe encwadini
- Okungenzeka
- Ukufunda okujulile
- I-L1 ejwayelekile
- Ukuthuthukisa
- Ukucubungula umbhalo
- Izinhlelo zokusebenza ze-Computer Vision
- Izinhlelo zokusebenza zamarobhothi
11. Izici Zokufunda Kwezibalo
Ngohlaka lwayo lomqondo kanye nezifundo ezahlukahlukene, le ncwadi yokufunda yomshini ivamise ukuvunywa kulo mkhakha.
Leli bhuku lingasetshenziswa njengereferensi yanoma ubani odinga ukubhulasha ngezihloko ezifana namanethiwekhi e-neural namasu okuhlola kanye nesingeniso esilula sokufunda ngomshini.
Incwadi icindezela umfundi ngokunamandla ukuthi azenzele ezabo izivivinyo nophenyo ngaso sonke isikhathi, ikwenze kube wusizo ekuthuthukiseni amakhono nelukuluku elidingekayo ukuze enze intuthuko efanelekile kumthamo wokufunda womshini noma umsebenzi.
Kuyithuluzi elibalulekile lezazi zezibalo nanoma ubani onentshisekelo yokumbiwa kwedatha ebhizinisini noma kwisayensi. Qiniseka ukuthi uyayiqonda i-algebra yomugqa okungenani ngaphambi kokuqala le ncwadi.
Izihloko ezifakwe encwadini
- Ukufunda okugadiwe (ukubikezela) kokufunda okungagadiwe
- Ama-Neural amanethiwekhi
- Imishini yokusekela i-vector
- Izihlahla zokuhlukanisa
- Ukuthuthukisa ama-algorithms
12. Ukuqashelwa Kwephethini Nokufunda Ngomshini
Imihlaba yokubonwa kwephethini nokufunda komshini kungahlolwa kahle kuleli bhuku. Indlela ye-Bayesian yokuqashelwa kwephethini yethulwa kulokhu kushicilelwa.
Ngaphezu kwalokho, le ncwadi ihlola izifundo eziyinselele ezidinga ukuqonda okusebenzayo kwe-multivariate, isayensi yedatha, kanye ne-algebra yomugqa obalulekile.
Ekufundeni komshini namathuba, ibhuku lesithenjwa linikeza izahluko ezinamaleveli aqina kancane kancane okuba yinkimbinkimbi ngokusekelwe kumathrendi kumadathasethi. Izibonelo ezilula zinikezwa ngaphambi kwesingeniso esijwayelekile sokubonwa kwephethini.
Incwadi inikeza amasu okulinganisa okulinganiselwe, okuvumela ukulinganiselwa okusheshayo ezimeni lapho izixazululo eziqondile zingenakwenzeka. Azikho ezinye izincwadi ezisebenzisa amamodeli ezithombe ukuchaza amathuba okusabalalisa, kodwa kunjalo.
Izihloko ezifakwe encwadini
- Izindlela zaseBayesia
- I-algorithms yokucabanga elinganiselwe
- Amamodeli amasha asuselwe kuma-kernel
- Isingeniso sethiyori yamathuba ayisisekelo
- Isingeniso sokubonwa kwephethini nokufunda komshini
13. Izisekelo Zokufunda Ngomshini kusukela ku-Predictive Data Analytics
Uma wazi kahle izinto eziyisisekelo zokufunda komshini futhi ufuna ukudlulela kuzibalo zedatha ezibikezelayo, leli yincwadi yakho!!! Ngokuthola amaphethini kumadathasethi amakhulu, Ukufunda Ngomshini kungasetshenziswa ukuthuthukisa amamodeli okuqagela.
Le ncwadi ihlola ukusetshenziswa kokusetshenziswa kwe-ML I-Predictive Data Analytics ngokujulile, okuhlanganisa kokubili imigomo yetiyori kanye nezibonelo zangempela.
Naphezu kweqiniso lokuthi isihloko esithi "Izisekelo Zokufunda Ngomshini Zokuhlaziywa Kwedatha Okubikezelwayo" siwumlomo, leli bhuku lizoveza uhambo lwe-Predictive Data Analytics ukusuka kudatha kuya ekuqondeni kuya esiphethweni.
Iphinde idingide izindlela ezine zokufunda komshini: ukufunda okusekelwe olwazini, ukufunda okusekelwe ekufananeni, ukufunda okusekelwe emathubeni, nokufunda okusekelwe emaphutheni, ngayinye enencazelo engeyona yobuchwepheshe elandelwa amamodeli ezibalo nama-algorithms anezibonelo.
Izihloko Ezihlanganiswe encwadini
- Ukufunda okusekelwe olwazini
- Ukufunda okusekelwe okufanayo
- Ukufunda okusekelwe emathubeni
- Ukufunda okusekelwe emaphutheni
14. I-Applied Predictive Modeling
I-Applied Predictive Modeling ihlola yonke inqubo yokubikezela yokubikezela, eqala ngezigaba ezibucayi zokucubungula ngaphambilini idatha, ukuhlukaniswa kwedatha, nezisekelo zokushuna amamodeli.
Umsebenzi ube usunikeza izincazelo ezicacile zezindlela ezihlukahlukene ezivamile nezakamuva zokuhlehla nokuhlela, ngokugxila ekuboniseni nasekuxazululeni izinselele zedatha zomhlaba wangempela.
Umhlahlandlela ubonisa zonke izici zenqubo yokumodela ngezandla ezimbalwa, izibonelo zomhlaba wangempela, futhi isahluko ngasinye sihlanganisa ikhodi engu-R ebanzi yesigaba ngasinye senqubo.
Lo mthamo wezinhloso eziningi ungasetshenziswa njengesethulo kumamodeli abikezelayo kanye nayo yonke inqubo yokumodela, njengenkomba yenkomba yodokotela, noma njengombhalo wezifundo zokubikezela ezithuthukisiwe zeziqu eziphakeme noma zeziqu.
Izihloko ezifakwe encwadini
- Indlela yokuhlehla
- Indlela yokuhlukanisa
- Ama-algorithms e-ML ayinkimbinkimbi
15. Ukufunda Ngomshini: Ubuciko Nesayensi Yama-algorithms Enza Umuzwa Wedatha
Uma ungumuntu ophakathi nendawo noma uchwepheshe wokufunda ngomshini futhi ufuna ukubuyela “kuzinto eziyisisekelo,” le ncwadi ingeyakho! Ikhokha isikweletu esigcwele ngobunkimbinkimbi nokujula koMshini omkhulu kuyilapho ingalahlekelwa umbono wezimiso zayo ezihlanganisayo (impumelelo impela!).
Ukufunda Ngomshini: Ubuciko Nesayensi Yama-algorithms afaka phakathi izifundo zezenzakalo ezikhulayo zobunzima obukhulayo, kanye nezibonelo eziningi nezithombe (ukugcina izinto ezithakazelisayo!).
Le ncwadi iphinde ihlanganise anhlobonhlobo amamodeli anengqondo, ejiyomethri, nezibalo, kanye nezihloko eziyinkimbinkimbi nezinoveli ezifana ne-matrix factorization kanye nokuhlaziywa kwe-ROC.
Izihloko ezifakwe encwadini
- Yenza lula ama-algorithms wokufunda komshini
- Imodeli enengqondo
- Imodeli yeJiyomethri
- Imodeli yesitatimende
- Ukuhlaziywa kwe-ROC
16. Ukumbiwa Kwedatha: Amathuluzi Okufunda Omshini Angokoqobo Namasu
Ngokusebenzisa izindlela ezisuka ocwaningweni lwezinhlelo zesizindalwazi, ukufunda ngomshini, nezibalo, amasu okumba idatha asenza sikwazi ukuthola amaphethini enanini elikhulu ledatha.
Kufanele uthole ibhuku elithi Imayini Yedatha: Amathuluzi Okufunda Omshini Awusizo Namasu uma udinga ukufunda amasu okumba idatha ikakhulukazi noma uhlele ukufunda ukufunda ngomshini ngokuvamile.
Incwadi engcono kakhulu yokufunda ngomshini igxile kakhulu ohlangothini lwayo lobuchwepheshe. Icubungula ngokwengeziwe ubunkimbinkimbi bobuchwepheshe bokufunda komshini, namasu okuqoqa idatha nokusebenzisa okokufaka okuhlukahlukene kanye nemiphumela ukuze kwahlulele imiphumela.
Izihloko ezifakwe encwadini
- Amamodeli alayini
- Ukuhlanganisa
- Ukumodela kwezibalo
- Ukubikezela ukusebenza
- Ukuqhathanisa izindlela zokumbiwa kwedatha
- Ukufunda okusekelwe emibonweni
- Ukumelwa kolwazi namaqoqo
- Izindlela zendabuko nesimanje zokumbiwa kwedatha
17. I-Python yokuhlaziywa kwedatha
Ikhono lokuhlola idatha esetshenziswa ekufundeni komshini yikhono elibaluleke kakhulu usosayensi wedatha okufanele abe nalo. Ngaphambi kokwenza imodeli ye-ML ekhiqiza isibikezelo sezulu esinembile, iningi lomsebenzi wakho lizobandakanya ukuphatha, ukucubungula, ukuhlanza, nokuhlola idatha.
Udinga ukujwayelana nezilimi zokuhlela ezifana nePandas, NumPy, Ipython, nezinye ukuze wenze ukuhlaziya idatha.
Uma ufuna ukusebenza kusayensi yedatha noma ukufunda komshini, kufanele ube nekhono lokukhohlisa idatha.
Kufanele nakanjani ufunde incwadi Python for Data Analysis kuleli cala.
Izihloko ezifakwe encwadini
- Okubalulekile Ama-Libraries asePython
- AmaPanda Athuthukile
- Izibonelo Zokuhlaziywa Kwedatha
- Ukuhlanzwa Kwedatha kanye Nokulungiselela
- Izindlela Zezibalo Nezibalo
- Ukufingqa kanye neComputing Izibalo Ezichazayo
18. Ukucubungula Ulimi Lwemvelo ngePython
Isisekelo sezinhlelo zokufunda zomshini ukucutshungulwa kolimi lwemvelo.
Incwadi ethi Natural Language Processing with Python ikuyala ukuthi ungasebenzisa kanjani i-NLTK, iqoqo elithandwa kakhulu lamamojula wePython namathuluzi okucubungula ulimi lwemvelo olungokomfanekiso nezibalo lwesiNgisi ne-NLP ngokujwayelekile.
Incwadi ethi Natural Language Processing ngePython inikeza izindlela ezisebenzayo zePython ezibonisa i-NLP ngendlela emfushane, esobala.
Abafundi banokufinyelela kumadathasethi anezichasiselo ezinhle zokubhekana nedatha engahlelekile, ukwakheka kolimi lombhalo, nezinye izici ezigxile ku-NLP.
Izihloko ezifakwe encwadini
- Lusebenza kanjani ulimi lwabantu?
- Izakhiwo zedatha yolimi
- Ikhithi yamathuluzi yolimi lwemvelo (NLTK)
- Ukuhlaziya nokuhlaziya kwe-semantic
- Izizindalwazi zolimi ezidumile
- Hlanganisa amasu kusuka ukuhlakanipha okungekhona okwangempela kanye nezezilimi
19. Ukuhlela Ukuhlanganyela Okuhlangene
I-Programming Collective Intelligence ka-Toby Segaran, ethathwa njengenye yezincwadi ezinkulu kakhulu zokuqala ukuqonda ukufundwa komshini, yabhalwa ngo-2007, iminyaka ngaphambi kokuba isayensi yedatha nokufunda ngomshini kufinyelele isikhundla sakho samanje njengezindlela zochwepheshe ezihamba phambili.
Le ncwadi isebenzisa iPython njengendlela yokusabalalisa ubuchwepheshe bayo kubabukeli bayo. I-Programming Collective Intelligence ingaphezulu kwebhukwana lokusetshenziswa kwe-ml kunesethulo sokufunda ngomshini.
Incwadi ihlinzeka ngolwazi lokuthuthukisa ama-algorithms e-ML asebenzayo okuqoqa idatha kusuka kuzinhlelo zokusebenza, ukuhlela ukuthola idatha kumawebhusayithi, kanye nokwengeza idatha eqoqiwe.
Isahluko ngasinye sihlanganisa imisebenzi yokwandisa ama-algorithm okuxoxiwe ngawo kanye nokuthuthukisa ukusebenza kwawo.
Izihloko ezifakwe encwadini
- Ukuhlunga kwe-Bayesia
- Imishini yokusekela i-vector
- Ama-algorithms enjini yokusesha
- Izindlela zokwenza izibikezelo
- Amasu okuhlunga ngokubambisana
- I-non-negative matrix factorization
- Ubuhlakani obuguqukayo bokuxazulula izinkinga
- Izindlela zokuthola amaqembu noma amaphethini
20. Ukufunda Okujulile (i-Adaptive Computation kanye nochungechunge lokufunda ngomshini)
Njengoba sonke sazi, ukufunda okujulile kuwuhlobo oluthuthukisiwe lokufunda komshini olwenza amakhompyutha afunde ekusebenzeni kwangaphambilini kanye nenani elikhulu ledatha.
Ngenkathi usebenzisa izindlela zokufunda zomshini, udinga ukwazi nezimiso zokufunda ezijulile. Le ncwadi, ethathwa njengeBhayibheli lemfundo ejulile, izoba usizo kakhulu kulesi simo.
Ochwepheshe abathathu bokufunda okujulile bahlanganisa izihloko eziyinkimbinkimbi kakhulu ezigcwele izibalo kanye namamodeli ajulile akhiqizayo kule ncwadi.
Ngokuhlinzeka ngesisekelo sezibalo nesomqondo, umsebenzi uxoxa ngemibono efanelekile ku-algebra yomugqa, ithiyori yamathuba, ithiyori yolwazi, ukubala kwezinombolo, nokufunda komshini.
Ihlola izinhlelo zokusebenza ezifana nokucutshungulwa kolimi lwemvelo, ukuqashelwa kwenkulumo, umbono wekhompuyutha, amasistimu wokuncoma ku-inthanethi, i-bioinformatics, nemidlalo yevidiyo futhi ichaza izindlela zokufunda ezijulile ezisetshenziswa ochwepheshe bemboni, njengamanethiwekhi e-feedforward ajulile, ukujwayela, nokwenza ngcono ama-algorithms, amanethiwekhi okuxhumana, kanye nendlela esebenzayo. .
Izihloko ezifakwe encwadini
- Ukubalwa kwamanani
- Ucwaningo Lokufunda Okujulile
- Amasu e-Computer Vision
- Amanethiwekhi e-Deep Feedforward
- Ukuthuthukisa Amamodeli Ajulile Wokuqeqesha
- Indlela Esebenzayo
- Ucwaningo Lokufunda Okujulile
Isiphetho
Amabhuku okufunda ngomshini aphezulu angama-20 afinyezwa kulolo hlu, ongalusebenzisa ukuze uthuthukise ukufunda komshini ngendlela othanda ngayo.
Uzokwazi ukwakha isisekelo esiqinile kulwazi lokufunda komshini kanye nelabhulali yesithenjwa ongayisebenzisa kaningi ngenkathi usebenza endaweni uma ufunda izinhlobonhlobo zalezi zincwadi zokufunda.
Uzogqugquzeleka ukuthi uqhubeke ufunda, uba ngcono, futhi ube nomthelela ngisho noma ufunda incwadi eyodwa nje.
Uma uzilungiselele futhi ukwazi ukuthuthukisa owakho ama-algorithms wokufunda komshini, khumbula ukuthi idatha ibaluleke kakhulu empumelelweni yephrojekthi yakho.
shiya impendulo