Kodi mungakonde kuyamba nawo makina kuphunzira?
Ndapanga phunziro losavuta komanso losavuta kwa oyamba kumene. Pamodzi, tikambirana njira zoyambira zophunzitsira makina ophunzirira makina.
Pamene ndikufotokozera njira zophunzitsira chitsanzo chimodzi ndi chimodzi, ndiperekanso chitsanzo choyambirira cha vuto la kuphunzira makina. Chifukwa chake, ngati mukufuna kutsatira, mutha kutsitsa chitsanzo ichi kuchokera pa izi kugwirizana.
Ichi ndi chitsanzo chabe cha deta chokuthandizani kuti muyambe kuphunzira pamakina.
Tili ndi zikhalidwe 18 za anthu amisinkhu yosiyana ndi amuna ndi akazi omwe nyimbo zomwe amakonda zimatanthauzidwa. Pogwiritsa ntchito, mawonekedwe a "zaka" ndi "jenda" tidzayesa kulingalira mtundu wa nyimbo womwe amakonda kwambiri.
Zindikirani: 1 ndi 0 amaperekedwa kwa amuna ndi akazi monga akazi ndi amuna mugululi.
Komabe, ngati simukufuna kutengera chitsanzo, ndi bwinonso. Ndikhala ndikufotokozera masitepe onsewa mwatsatanetsatane. Kotero, tiyeni tilowemo!
Zinthu Zoyamba Kudziwa
Tisanalowe masitepe ophunzitsira chitsanzo, tiyeni tifotokoze mfundo zina. Kuphunzira makina ndi nzeru zochita kupanga chilango chomwe chimayang'ana pakupanga ma algorithms omwe angaphunzire kuchokera ku data.
Kuti muchite izi, makina ophunzirira amaphunzitsidwa pa dataset yomwe imaphunzitsa chitsanzocho momwe mungalosere zolondola kapena gulu pazatsopano, zosadziwika kale.
Kotero, kodi zitsanzozi ndi ziti? A makina kuphunzira chitsanzo n'chimodzimodzi ndi njira imene kompyuta imagwiritsa ntchito kupanga zolosera kapena zosankha.
Chitsanzo, monga chophikira, chimatsatira ndondomeko ya malangizo kuti awunike deta ndi kupanga zolosera kapena ziganizo kutengera zomwe zapezeka mu data. Deta yochuluka yomwe chitsanzocho chimaphunzitsidwa, zolosera zake zimakhala zolondola kwambiri.
Ndi Zitsanzo Zotani Zomwe Tingaphunzitse?
Tiyeni tiwone mitundu yoyambira yophunzirira makina.
- Linear Regression: chitsanzo chomwe chimalosera kusinthika kwa chandamale kuchokera pamitundu imodzi kapena zingapo zolowetsa.
- Neural Networks: netiweki yama node olumikizidwa omwe angaphunzire kuzindikira machitidwe ovuta mu data.
- Mitengo Yopangira zisankho: njira yopangira zisankho yomangidwa panthambi zanthambi ngati-mwina.
- Clustering: gulu la zitsanzo zomwe zimagwirizanitsa mfundo zofananira potengera kufanana.
- Logistic Regression: chitsanzo cha zovuta zamagulu a binary pomwe chandamale chimakhala ndi zikhalidwe ziwiri.
- Mitengo Yopangira zisankho: njira yopangira zisankho yomangidwa panthambi zanthambi ngati-mwina.
- Random Forest: chitsanzo chophatikizana chopangidwa ndi mitengo yambiri yosankha. Amagwiritsidwa ntchito nthawi zambiri kuyika magulu ndi mapulogalamu obwerera.
- Oyandikana nawo a K-Nearest: chitsanzo chomwe chimalosera kusinthika kwa chandamale pogwiritsa ntchito ma data a k-apafupi kwambiri pamaphunzirowo.
Kutengera vuto lathu ndi dawunilodi, timasankha makina ophunzirira makina omwe akuyenerana ndi momwe tilili. Komabe, tibwereranso ku izi pambuyo pake. Tsopano, tiyeni tiyambe kuphunzitsa chitsanzo chathu. Ndikukhulupirira kuti mwatsitsa kale pepala ngati mungafune kutsatira chitsanzo chathu.
Komanso, ndikupangira kukhala Buku la Jupyter imayikidwa pamakina akomweko ndikuigwiritsa ntchito pama projekiti anu ophunzirira makina.
1: Fotokozani vuto
Gawo loyamba mu kuphunzitsa makina ophunzirira model ikufotokoza vuto lomwe liyenera kuthetsedwa. Izi zikuphatikizapo kusankha zosintha zomwe mukufuna kulosera (zomwe zimadziwika kuti chandamale) ndi zosintha zomwe zidzagwiritsidwe ntchito kupanga maulosi amenewo (odziwika ngati mawonekedwe kapena zolosera).
Muyeneranso kusankha mtundu wavuto lophunzirira makina lomwe mukuyesera kuthana nalo (kugawa, kubweza, kusanja, ndi zina zotero) ndi mtundu wanji wa deta yomwe mungafunikire kusonkhanitsa kapena kuphunzitsa chitsanzo chanu.
Mtundu wa mtundu womwe mumagwiritsa ntchito umatsimikiziridwa ndi mtundu wamavuto ophunzirira makina omwe mukufuna kuthetsa. Kugawikana, kutsika, ndi kusanjana ndi magulu atatu oyambirira a zovuta kuphunzira makina. Mukafuna kulosera zamitundu yosiyanasiyana, monga ngati imelo ndi sipamu kapena ayi, mumagwiritsa ntchito magulu.
Mukafuna kulosera zakusintha kosalekeza, monga mtengo wanyumba, mumagwiritsa ntchito regression. Clustering imagwiritsidwa ntchito kusonkhanitsa zinthu zofananira za data kutengera zomwe zimafanana.
Ngati tiyang'ana chitsanzo chathu; vuto lathu ndi kudziwa munthu amakonda nyimbo kalembedwe kuchokera jenda ndi zaka. Tigwiritsa ntchito gulu la anthu 18 pachitsanzo ichi komanso zambiri zazaka zawo, jenda, komanso nyimbo zomwe amakonda.
2. Konzani deta
Mutatha kufotokoza vuto, muyenera kukonzekera deta yophunzitsira chitsanzocho. Izi zikuphatikizapo kuyeretsa ndi kukonza deta. Chifukwa chake, titha kutsimikizira kuti ili m'mawonekedwe omwe a makina ophunzirira algorithm angagwiritse ntchito.
Izi zingaphatikizepo zinthu monga kufufuta zinthu zomwe zikusowa, kusintha deta yamagulu kukhala manambala, ndi kukulitsa kapena kukonzanso deta kuti zitsimikizire kuti zonse zili pamlingo wofanana.
Mwachitsanzo, umu ndi momwe mumachotsera zinthu zomwe zikusowa:
import pandas as pd
# Load the data into a pandas DataFrame
data = pd.read_csv('data.csv')
# Check for missing values
print(data.isnull().sum())
# Drop rows with missing values
data.dropna(inplace=True)
# Check that all missing values have been removed
print(data.isnull().sum())
Chidziwitso chaching'ono: Pamzere o "import pandas as pd",
timalowetsa laibulale ya Pandas ndikuipatsa dzina loti "pd" kuti ikhale yosavuta kufotokozera ntchito zake ndi zinthu pambuyo pake mu code.
Pandas ndi gawo lodziwika bwino la Python pakuwongolera ndi kusanthula deta, makamaka pogwira ntchito ndi data yokhazikika kapena tabular.
Mu chitsanzo chathu chodziwitsa mitundu ya nyimbo. Tibweretsa kaye deta. Ndachitcha music.csv, komabe, mutha kutchula momwe mungafune.
Kuti tikonzekeretse deta yophunzitsira makina ophunzirira makina, timawagawa m'magulu (zaka ndi jenda) ndi zolinga (mtundu wanyimbo).
Tidzagawanso zidziwitsozo mu 80:20 seti yophunzitsira ndi kuyesa kuti tiwone momwe chitsanzo chathu chimagwirira ntchito ndikupewa kuchulukitsitsa.
# Import necessary libraries
import pandas as pd
from sklearn.model_selection import train_test_split
# Load data from CSV file/code>
music_data = pd.read_csv('music.csv')
# Split data into features and target
X = music_data.drop(columns=['genre'])
y = music_data['genre']
# Split data into training and testing sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
3. Sankhani makina ophunzirira makina.
Pambuyo pokonzekera deta, muyenera kusankha chitsanzo chophunzirira makina chomwe chikugwirizana ndi ntchito yanu.
Pali ma aligorivimu angapo oti musankhe, monga mitengo yaziganizo, kusinthika kwazinthu, makina othandizira ma vector, ma neural network, ndi ena. Algorithm yomwe mwasankha idzatsimikiziridwa ndi mtundu wankhani yomwe mukuyesera kuyankha, mtundu wa data yomwe muli nayo, ndi zosowa zanu.
Tigwiritsa ntchito gulu lamitengo pachitsanzo ichi chifukwa tikugwira ntchito ndi vuto lamagulu (kuneneratu zamagulu).
# Import necessary libraries
from sklearn.tree import DecisionTreeClassifier
Nayi chithunzithunzi cha momwe Decision Tree Classifier imagwirira ntchito:
4. Phunzitsani chitsanzo
Mutha kuyamba kuphunzitsa chitsanzocho mutasankha njira yovomerezeka yophunzirira makina. Izi zimaphatikizapo kugwiritsa ntchito zomwe zidapangidwa m'mbuyomu kuphunzitsa ma algorithm amomwe mungalosere pazatsopano, zomwe sizinawonekere.
Algorithm idzasintha magawo ake amkati panthawi yophunzitsira kuti achepetse kusiyana pakati pa zomwe zanenedweratu ndi zomwe zili muzolemba zamaphunziro. Kuchuluka kwa deta yomwe imagwiritsidwa ntchito pophunzitsa, komanso magawo ake enieni a algorithm, zonse zitha kukhala ndi zotsatira pakulondola kwa chitsanzo chotsatira.
Muchitsanzo chathu chenicheni, popeza tasankha njira, titha kuphunzitsa chitsanzo chathu ndi data yophunzitsira.
# Train the decision tree classifier
model = DecisionTreeClassifier()
model.fit(X_train, y_train)
5. Unikani chitsanzo
Chitsanzocho chikaphunzitsidwa, chiyenera kuyesedwa pa deta yatsopano kuti zitsimikizire kuti ndizolondola komanso zodalirika. Izi zikuphatikizapo kuyesa chitsanzocho ndi deta yomwe sinagwiritsidwe ntchito panthawi ya maphunziro ndi kufananitsa zikhulupiliro zake ndi zofunikira zenizeni zomwe zili mu data yoyesera.
Ndemangayi ingathandize kuzindikira zolakwika zilizonse zachitsanzo, monga kuwonjezereka kapena kuperewera, ndipo zingayambitse kukonza bwino komwe kungafunike.
Pogwiritsa ntchito deta yoyesera, tidzayesa kulondola kwa chitsanzo chathu.
# Import necessary libraries
from sklearn.metrics import accuracy_score
# Predict the music genre for the test data
predictions = model.predict(X_test)
# Evaluate the model's accuracy
accuracy = accuracy_score(y_test, predictions)
print("Accuracy: ", accuracy)
Kulondola kwabwino sikuli koyipa kwambiri pakadali pano. 🙂 Kuti muwongolere kulondola kwanu, mutha kuyeretsa zambiri nthawi zonse kapena kuyesa mitundu yosiyanasiyana yophunzirira pamakina kuti muwone yemwe amapereka kwambiri.
6. Konzani bwino chitsanzo
Ngati luso lachitsanzolo silikukwanira, mutha kuyikonza bwino posintha magawo osiyanasiyana a algorithm kapena kuyesa ma algorithms atsopano kwathunthu.
Njirayi ingaphatikizepo kuyesa njira zina zophunzirira, kusintha makonda, kapena kusintha kuchuluka kapena kukula kwa zigawo zobisika mu neural network.
7. Gwiritsani ntchito chitsanzo
Mukangokondwera ndi machitidwe achitsanzo, mukhoza kuyamba kugwiritsa ntchito kupanga zolosera pazatsopano zatsopano.
Izi zitha kuphatikizira kudyetsa deta yatsopano muchitsanzocho ndikugwiritsa ntchito zomwe zaphunziridwa zachitsanzo kupanga zolosera pa datayo, kapena kuphatikiza chitsanzocho munjira yotakata kapena dongosolo.
Titha kugwiritsa ntchito chitsanzo chathu kupanga zolosera pazatsopano pambuyo poti tasangalala ndi kulondola kwake. Mukhoza kuyesa zosiyana siyana za jenda ndi zaka.
# Test the model with new data
new_data = [[25, 1], [30, 0]]
predictions = model.predict(new_data)
print("Predictions: ", predictions)
Womba mkota
Tamaliza kuphunzitsa chitsanzo chathu choyamba cha makina ophunzirira.
Ndikukhulupirira kuti mwapeza zothandiza. Tsopano mutha kuyesa kugwiritsa ntchito mitundu yosiyanasiyana yophunzirira makina monga Linear Regression kapena Random Forest.
Pali ma dataset ambiri ndi zovuta mu Chitani ngati mukufuna kukonza zolemba zanu komanso kumvetsetsa kwa kuphunzira pamakina.
Siyani Mumakonda