Ayyuka 14 na buɗe tushen don haɓaka ƙwarewar Kimiyyar Bayanai (sauki, al'ada, mai wuya)

Kimiyyar Bayanai don Mafari

1. Nazari na Ji (Sentiment Analysis Ta Rubutu)

Duba cikakken aiwatar da aikin Kimiyyar Bayanai ta amfani da lambar tushe - Ayyukan Binciken Sentiment a cikin R.

Binciken Sentiment shine nazarin kalmomi don tantance ra'ayi da ra'ayi, wanda zai iya zama mai kyau ko mara kyau. Wannan nau'i ne na rarrabuwa wanda azuzuwan na iya zama binary (tabbatacce da mara kyau) ko jam'i (mai farin ciki, fushi, bakin ciki, rashin tausayi ...). Za mu aiwatar da wannan aikin Kimiyyar Bayanai a cikin R kuma za mu yi amfani da bayanan da ke cikin kunshin "janeaustenR". Za mu yi amfani da ƙamus na gaba ɗaya kamar AFINN, Bing da loughran, mu yi haɗin ciki, kuma a ƙarshe za mu ƙirƙiri kalmar girgije don nuna sakamakon.

Harshe: R
Saitin bayanai/ Kunshin: janeaustenR

An fassara labarin tare da tallafin EDISON Software, wanda yana yin ɗakuna masu dacewa na kama-da-wane don shagunan iri iri-iriKuma gwada software.

2. Gano Labaran Karya

Ɗauki ƙwarewar ku zuwa mataki na gaba ta hanyar aiki akan aikin Kimiyyar Bayanai don masu farawa - gano labaran karya da Python.

Labaran karya labarai ne na karya da ake yadawa ta kafafen sada zumunta da sauran kafafen sadarwa na yanar gizo don cimma manufofin siyasa. A cikin wannan ra'ayin aikin Kimiyyar Bayanai, za mu yi amfani da Python don gina ƙirar da za ta iya tantance daidai ko labarin gaskiya ne ko na karya. Za mu ƙirƙiri TfidfVectorizer kuma mu yi amfani da PassiveAggressiveClassifier don rarraba labarai zuwa "hakikanin" da "karya". Za mu yi amfani da ma'aunin bayanai na siffar 7796×4 kuma mu gudanar da komai a cikin Jupyter Lab.

Harshe: Python

Saitin bayanai/ Kunshin: labarai.csv

3. Gano Cutar Parkinson

Ci gaba tare da ra'ayin Project Science Project - gano cutar Parkinson ta amfani da XGBoost.

Mun fara amfani da Kimiyyar Bayanai don inganta kiwon lafiya da ayyuka - idan za mu iya hasashen cuta a farkon matakin, to za mu sami fa'idodi da yawa. Don haka, a cikin wannan tunanin aikin Kimiyyar Bayanai, za mu koyi yadda ake gano cutar Parkinson ta amfani da Python. Yana da neurodegenerative, cututtuka na ci gaba na tsarin kulawa na tsakiya wanda ke shafar motsi kuma yana haifar da rawar jiki da taurin kai. Yana shafar kwayoyin halitta masu samar da dopamine a cikin kwakwalwa, kuma a kowace shekara, yana shafar fiye da mutane miliyan 1 a Indiya.

Harshe: Python

Saitin bayanai/ Kunshin: UCI ML Parkinsons dataset

Ayyukan Kimiyyar Bayanai na matsakaicin rikitarwa

4. Gane Ƙaunar Magana

Duba cikakken aiwatar da aikin misalin Kimiyyar Bayanai - gane magana ta amfani da Librosa.

Yanzu bari mu koyi yadda ake amfani da ɗakunan karatu daban-daban. Wannan aikin Kimiyyar Bayanai yana amfani da librosa don gane magana. SER shine tsari na gano motsin zuciyar ɗan adam da jihohi masu tasiri daga magana. Tun da muke amfani da sauti da sauti don bayyana motsin rai tare da muryoyin mu, SER ya dace. Amma tunda motsin rai na zahiri ne, bayanin sautin aiki ne mai wahala. Za mu yi amfani da mfcc, chroma da ayyukan mel kuma za mu yi amfani da saitin bayanan RAVDESS don fahimtar motsin rai. Za mu ƙirƙiri mai rarraba MLPC don wannan ƙirar.

Harshe: Python

Saitin bayanai/ Kunshin: Bayanan Bayani na RAVDESS

5. Gano Jinsi da Shekaru

Buga ma'aikata da sabon aikin Kimiyyar Bayanai - ƙayyade jinsi da shekaru ta amfani da OpenCV.

Wannan Kimiyyar Bayanai ce mai ban sha'awa tare da Python. Yin amfani da hoto ɗaya kawai, za ku koyi hasashen jinsi da shekarun mutum. A cikin wannan za mu gabatar muku da Vision Vision da ka'idodinta. Za mu yi gini convolutional jijiya cibiyar sadarwa kuma za su yi amfani da samfura waɗanda Tal Hassner da Gil Levy suka horar akan bayanan Adience. A hanya za mu yi amfani da wasu fayilolin .pb, .pbtxt, .prototxt da .caffemodel.

Harshe: Python

Saitin bayanai/ Kunshin: Adience

6. Uber Data Analysis

Duba cikakken aiwatar da aikin Kimiyyar Bayanai tare da lambar tushe - Uber Data Analysis Project a cikin R.

Wannan aikin hangen nesa ne na bayanai tare da ggplot2 wanda a cikinsa za mu yi amfani da R da ɗakunan karatu tare da nazarin sigogi daban-daban. Za mu yi amfani da bayanan Uber Pickups New York City kuma mu ƙirƙira abubuwan gani don firam ɗin lokaci daban-daban na shekara. Wannan yana gaya mana yadda lokaci ke shafar tafiyar abokin ciniki.

Harshe: R

Saitin bayanai/ Kunshin: Uber Pickups a cikin New York City dataset

7. Gano Direba Direba

Haɓaka ƙwarewar ku ta yin aiki akan Babban Aikin Kimiyyar Bayanai - tsarin gano bacci tare da OpenCV & Keras.

Tuki a nutse yana da hatsarin gaske, kuma kusan hatsarurruka dubu na faruwa a duk shekara sakamakon barcin da direbobi ke yi a yayin tuki. A cikin wannan aikin Python, za mu ƙirƙiri tsarin da zai iya gano direbobin barci da kuma faɗakar da su da siginar sauti.

Ana aiwatar da wannan aikin ta amfani da Keras da OpenCV. Za mu yi amfani da OpenCV don gano fuska da ido kuma tare da Keras za mu rarraba yanayin ido (Buɗe ko Rufe) ta amfani da dabarun hanyar sadarwa mai zurfi.

8. Chatbot

Ƙirƙiri Chatbot tare da Python kuma ɗauki mataki gaba a cikin aikinku - Chatbot tare da NLTK & Keras.

Chatbots wani bangare ne na kasuwanci. Yawancin kasuwancin dole ne su ba da sabis ga abokan cinikinsu kuma yana ɗaukar ƙarfi, lokaci da ƙoƙari don yi musu hidima. Chatbots na iya sarrafa yawancin hulɗar abokin ciniki ta hanyar amsa wasu tambayoyin gama-gari waɗanda abokan ciniki ke yi. Akwai ainihin nau'ikan chatbots guda biyu: takamaiman yanki da Buɗe-domain. Ana yawan amfani da ƙayyadaddun taswirar yanki don magance takamaiman matsala. Don haka, kuna buƙatar keɓance shi don yin aiki yadda ya kamata a fagen ku. Za a iya yi wa budadden yanki chatbots kowace tambaya, don haka horar da su yana buƙatar adadi mai yawa na bayanai.

Saitin bayanai: Intents json fayil

Harshe: Python

Advanced Data Science ayyukan

9. Hoto Generator

Duba cikakken aiwatar da aikin tare da lambar tushe - Generator Taken Hoto tare da CNN & LSTM.

Bayyana abin da ke cikin hoto aiki ne mai sauƙi ga ɗan adam, amma ga kwamfutoci, hoto kawai jerin lambobi ne waɗanda ke wakiltar ƙimar launi na kowane pixel. Wannan aiki ne mai wahala ga kwamfutoci. Fahimtar abin da ke cikin hoto sannan ƙirƙirar bayanin a cikin harshe na halitta (kamar Ingilishi) wani aiki ne mai wahala. Wannan aikin yana amfani da dabarun ilmantarwa mai zurfi wanda a cikinsa muke aiwatar da hanyar sadarwa ta Convolutional Neural Network (CNN) tare da Recurrent Neural Network (LSTM) don ƙirƙirar janareta bayanin hoto.

Saitin bayanai: Farashin 8K

Harshe: Python

Tsarin: Keras

10. Gano Zamban Katin Kiredit

Yi iya ƙoƙarinku yayin aiki akan ra'ayin aikin Kimiyyar Bayanai - gano zamba ta hanyar amfani da na'ura koyo.

A yanzu kun fara fahimtar dabaru da dabaru. Bari mu ci gaba zuwa wasu ci-gaban ayyukan kimiyyar bayanai. A cikin wannan aikin za mu yi amfani da harshen R tare da algorithms kamar yanke shawara bishiyoyi, koma-bayan dabaru, hanyoyin sadarwa na jijiyoyi na wucin gadi da haɓaka mai ƙima. Za mu yi amfani da bayanan ma'amalar katin don rarraba ma'amalar katin kiredit a matsayin yaudara ko na gaske. Za mu zaɓi samfuri daban-daban a gare su kuma mu gina masu lanƙwasa.

Harshe: R

Saitin bayanai/ Kunshin: Saitin bayanan Ma'amalolin Kati

11. Tsarin Shawarar Fim

Yi nazarin aiwatar da mafi kyawun aikin Kimiyyar Bayanai tare da lambar tushe - Tsarin Shawarar Fim a cikin harshen R

A cikin wannan aikin Kimiyyar Bayanai, za mu yi amfani da R don aiwatar da shawarwarin fim ɗin ta hanyar koyon inji. Tsarin shawarwarin yana aika shawarwari ga masu amfani ta hanyar tsarin tacewa bisa abubuwan da sauran masu amfani suka zaɓa da tarihin bincike. Idan A da B suna son Gida Kadai, kuma B suna son 'Yan Mata Ma'ana, to kuna iya ba da shawarar A - za su iya son shi ma. Wannan yana ba abokan ciniki damar yin hulɗa tare da dandamali.

Harshe: R

Saitin bayanai/ Kunshin: MovieLens dataset

12. Rarraba Abokin Ciniki

Buga ma'aikata aiki tare da aikin Kimiyyar Bayanai (ciki har da lambar tushe) - Rarraba abokin ciniki ta amfani da koyo na inji.

Bangaren mai siye sanannen aikace-aikace ne ilmantarwa mara kulawa. Yin amfani da tari, kamfanoni suna gano sassan abokan ciniki don ƙaddamar da yuwuwar tushen mai amfani. Suna rarraba abokan ciniki zuwa rukuni bisa ga halaye na gama gari kamar jinsi, shekaru, sha'awa da halaye na kashe kuɗi ta yadda za su iya tallata samfuransu yadda ya kamata ga kowane rukuni. Za mu yi amfani K- yana nufin tari, da kuma hangen nesa rarraba ta jinsi da shekaru. Daga nan za mu yi nazarin kuɗin shiga na shekara-shekara da matakan kashe kuɗi.

Harshe: R

Saitin bayanai/ Kunshin: Mall_Customers dataset

13. Rarraba Ciwon Daji

Duba cikakken aiwatar da aikin Kimiyyar Bayanai a Python - Rarraba kansar nono ta amfani da zurfafa ilmantarwa.

Idan muka dawo kan gudunmawar likitanci na kimiyyar bayanai, bari mu koyi yadda ake gano kansar nono ta amfani da Python. Za mu yi amfani da saitin bayanai na IDC_na yau da kullun don gano ciwon daji na ductal, mafi yawan nau'in ciwon nono. Yana tasowa a cikin magudanar madara, yana shiga cikin fibrous ko mai kitse a wajen bututun nono. A cikin wannan ra'ayin aikin kimiyyar tattara bayanai za mu yi amfani da shi Jin Ilimi da ɗakin karatu na Keras don rarrabawa.

Harshe: Python

Saitin bayanai/ Kunshin: IDC_na yau da kullun

14. Gane Alamomin Tafiya

Samun daidaito a cikin fasahar tuƙi tare da aikin Kimiyyar Bayanai Gane alamar zirga-zirga ta amfani da CNN bude tushen.

Alamun hanya da dokokin zirga-zirga suna da matukar mahimmanci ga kowane direba don guje wa haɗari. Don bin ƙa'idar, da farko kuna buƙatar fahimtar yadda alamar hanya take. Dole ne mutum ya koyi duk alamun hanya kafin a ba shi lasisin tuka kowace mota. Amma yanzu adadin motocin masu cin gashin kansu na karuwa, kuma nan gaba kadan mutum ba zai sake tuka mota da kansa ba. A cikin aikin Gane Alamar Hanya, zaku koyi yadda shirin zai iya gane nau'in alamun hanya ta hanyar ɗaukar hoto azaman shigarwa. Ana amfani da saitin Gane Alamar Traffic ta Jamus (GTSRB) don gina hanyar sadarwa mai zurfi don gane ajin da alamar zirga-zirga ta ke. Muna kuma ƙirƙirar GUI mai sauƙi don yin hulɗa tare da aikace-aikacen.

Harshe: Python

Saitin bayanai: GTSRB (Gane Alamar Traffic ta Jamus)

Kara karantawa

• 52 datasets don ayyukan horo
• Dojo na gaba-gaba: ayyukan don horar da ƙwarewar haɓaka (sabbi 5 + 43 tsoho)
• Top 12 mafi ban sha'awa IT tsauri infographics

source: www.habr.com