O ka rata ho qala ka ho ithuta mochine?
Ke thehile thuto e bonolo le e bonolo bakeng sa ba qalang ka ho feletseng. Hammoho, re tla fetela mehatong ea mantlha ea ho koetlisa mohlala oa ho ithuta mochini.
Ha ke ntse ke hlalosa mehato ea ho koetlisa mohlala ka bonngoe, ke tla boela ke fane ka mohlala oa motheo oa bothata ba ho ithuta ka mochine. Kahoo, haeba u batla ho latela, u ka khoasolla sampole ena ea data ho tsoa ho sena link tsa.
Ona ke mohlala feela oa pokello ea lintlha ho u thusa ho qala ka ho ithuta ka mochini.
Re na le litekanyetso tse 18 tsa batho ba lilemo tse fapaneng le ba bong bo fapaneng ba nang le 'mino oo ba o ratang haholo o hlalosoang. Ka ho sebelisa, likarolo tsa "lilemo" le "bong" re tla leka ho hakanya hore na ke mofuta ofe oa 'mino oo ba o ratang haholo.
Tlhokomeliso: 1 le 0 li fuoa banna le basali sebakeng sena sa data.
Leha ho le joalo, haeba u sa batle ho latela mohlala, ho boetse ho nepahetse. Ke tla be ke hlalosa mehato ena kaofela ka botlalo. Kahoo, ha re ikakhele ka setotsoana!
Lintho Tsa Pele Tseo U Lokelang ho li Tseba
Pele re kena mehatong ea ho koetlisa mohlala, a re hlakiseng lintlha tse ling. Ho ithuta ka mochine ke bohlale ba maiketsetso taeo e shebaneng le ho hlahisa li-algorithms tse ka ithutang ho data.
Ho etsa sena, mefuta ea ho ithuta ka mochini e koetlisetsoa pokellong ea data e rutang mohlala ho etsa likhakanyo tse nepahetseng kapa tlhophiso ho data e ncha, e neng e sa tsejoe pele.
Joale, mehlala ee ke efe? A mochini oa ho ithuta oa mechini e tšoana le risepe eo komporo e e sebelisang ho etsa likhakanyo tsa data kapa khetho.
Moetso, joalo ka risepe, o latela lethathamo la litaelo ho lekola datha le ho hlahisa likhakanyo kapa likahlolo tse ipapisitseng le lipaterone tse fumanehang ho data. Ha mohlala o ntse o koetlisoa ka boitsebiso bo bongata, ke moo likhakanyo tsa oona li nepahetseng haholoanyane.
Re ka Koetlisa Mehlala ea Mofuta Ofe?
Ha re boneng hore na mekhoa ea mantlha ea ho ithuta ka mochini ke efe.
- Linear Regression: mohlala o bolelang esale pele phetoho e tsoelang pele ea sepheo ho tsoa ho mofuta o le mong kapa ho feta.
- Neural Networks: marang-rang a li-node tse hokahaneng tse ka ithutang ho lemoha mekhoa e rarahaneng ho data.
- Lifate tsa Qeto: mokhoa oa ho etsa liqeto o hahiloeng holim'a ketane ea lipolelo tsa branching haeba-ho seng joalo.
- Clustering: sehlopha sa mehlala e hlophisang lintlha tse bapisoang tse ipapisitseng le ho tšoana.
- Logistic Regression: mohlala bakeng sa mathata a lihlopha tsa binary moo sepheo sa sepheo se nang le litekanyetso tse peli tse ka bang teng.
- Lifate tsa Qeto: mokhoa oa ho etsa liqeto o hahiloeng holim'a ketane ea lipolelo tsa branching haeba-ho seng joalo.
- Random Forest: mohlala o kopaneng o entsoeng ka lifate tse ngata tsa liqeto. Hangata li sebelisoa bakeng sa likopo tsa ho hlophisa le ho khutlisa.
- Baahelani ba K-Nearest: mohlala o bolelang esale pele phapang e reriloeng e sebelisa lintlha tse haufi tsa k-haufi sehlopheng sa koetliso.
Ho ipapisitsoe le bothata ba rona le datha, re etsa qeto ea hore na ke mofuta ofe oa ho ithuta oa mochini o loketseng maemo a rona haholo. Leha ho le joalo, re tla khutlela ho sena hamorao. Joale, ha re qaleng ho koetlisa mohlala oa rona. Ke tšepa hore u se u downloaded the setatata haeba u ka rata ho latela mohlala oa rona.
Hape, ke khothaletsa ho ba le Buka ea Jupyter e kentsoe mochining oa heno le ho e sebelisa bakeng sa merero ea hau ea ho ithuta ka mochini.
1: Hlalosa bothata
Mokhahlelo oa pele ho ho koetlisa ho ithuta ka mochini mohlala o hlalosa taba e lokelang ho rarolloa. Sena se kenyelletsa ho khetha mefuta-futa eo u lakatsang ho e bolela esale pele (e tsejoang ka hore ke sepheo se feto-fetohang) le mefuta e tla sebelisoa ho hlahisa likhakanyo tseo (tse tsejoang e le likarolo kapa li-predictors).
U lokela hape ho etsa qeto ea hore na u leka ho rarolla bothata ba mofuta ofe oa ho ithuta ka mochini (ho hlophisa, ho theola maemo, ho kopanya, joalo-joalo) le hore na u tla hloka data ea mofuta ofe ho e bokella kapa ho e fumana ho koetlisa mohlala oa hau.
Mofuta oa mofuta oo u o sebelisang o tla khethoa ke mofuta oa bothata ba ho ithuta ka mochini oo u ikemiselitseng ho o rarolla. Classification, regression, and clustering ke mekhahlelo e meraro ea mantlha ea mathata a ho ithuta ka mochini. Ha u batla ho bolela esale pele phapang ea likarolo, joalo ka hore na lengolo-tsoibila ke spam kapa che, u sebelisa classification.
Ha o lakatsa ho lepa phetoho e tsoelang pele, joalo ka theko ea ntlo, o sebelisa ho theoha. Clustering e sebelisoa ho kopanya lintlha tse bapisoang ho latela lintho tse tšoanang.
Haeba re sheba mohlala oa rona; phephetso ea rona ke ho tseba hore na motho o khetha mofuta ofe oa 'mino ho tloha ka bong le lilemo tsa bona. Re tla sebelisa sehlopha sa batho ba 18 ho etsa mohlala ona le lintlha tsa lilemo tsa bona, bong ba bona, le mofuta oa 'mino oo ba o ratang.
2. Lokisetsa lintlha
Ka mor'a hore u hlalose bothata, u tla hloka ho lokisa lintlha tsa ho koetlisa mohlala. Sena se kenyelletsa ho hloekisa le ho sebetsana le data. Kahoo, re ka etsa bonnete ba hore e ka sebopeho sa hore algorithm ea ho ithuta mochini ka sebelisa.
Sena se ka kenyelletsa mesebetsi e kang ho phumula boleng bo sieo, ho fetola lintlha tsa likarolo ho data tsa linomoro, le ho lelefatsa kapa ho tloaeleha ho etsa bonnete ba hore litšobotsi tsohle li lekana.
Ka mohlala, ena ke tsela eo u hlakolang litekanyetso tse sieo:
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())
Keletso e nyane: Moleng o "import pandas as pd",
re kenya laeborari ea Pandas mme re e abela lebitso "pd" ho etsa hore ho be bonolo ho supa mesebetsi le lintho tsa eona hamorao khoutu.
Pandas ke mojule o tsebahalang oa Python bakeng sa manollo le tlhahlobo ea data, haholo ha o sebetsa ka data e hlophisitsoeng kapa ea tabular.
Mohlala oa rona oa ho khetholla mefuta ea 'mino. Re tla qala ka ho kenya dataset. Ke e rehile music.csv, leha ho le joalo, u ka e reha ka tsela eo u batlang ka eona.
Ho lokisa lintlha tsa ho koetlisa mofuta oa ho ithuta mochini, re o arola ka litšoaneleho (lilemo le bong) le sepheo (mofuta oa 'mino).
Hape re tla arola lintlha ho li-sete tsa koetliso le liteko tsa 80:20 ho lekola ts'ebetso ea mohlala oa rona le ho qoba ho fetella.
# 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. Khetha mohlala oa ho ithuta mochine.
Ka mor'a hore u lokise lintlha, u tlameha ho khetha mohlala oa ho ithuta mochine o loketseng mosebetsi oa hau.
Ho na le li-algorithms tse 'maloa tseo u ka khethang ho tsona, joalo ka lifate tsa liqeto, ho fokotseha ha lintho, mechini ea li-vector ea tšehetso, marang-rang a neural, le tse ling. Algorithm eo u e khethang e tla khethoa ke mofuta oa bothata boo u lekang ho bo araba, mofuta oa data eo u nang le eona, le litlhoko tsa ts'ebetso ea hau.
Re tla sebelisa sehlopha sa sefate sa liqeto molemong oa mohlala ona hobane re sebetsana le bothata ba ho hlophisa (ho bolela esale pele lintlha tsa sehlopha).
# Import necessary libraries
from sklearn.tree import DecisionTreeClassifier
Mona ke pono ea kamoo Decision Tree Classifier e sebetsang kateng:
4. Koetlisa mohlala
U ka qala ho koetlisa mohlala ha u khethile algorithm e amohelehang ea ho ithuta ka mochini. Sena se kenyelletsa ho sebelisa lintlha tse hlahisitsoeng pele ho ruta algorithm mabapi le mokhoa oa ho bolela esale pele ka data e ncha, e neng e sa bonoe.
Algorithm e tla fetola mekhahlelo ea eona ea ka hare nakong ea koetliso ho fokotsa phapang pakeng tsa litekanyetso tsa eona tse boletsoeng esale pele le litekanyetso tsa sebele ho data ea koetliso. Bongata ba lintlha tse sebelisoang bakeng sa koetliso, hammoho le litekanyetso tse khethehileng tsa algorithm, kaofela li ka ba le phello ea ho nepahala ha mohlala oa sephetho.
Mohlala oa rona o ikhethileng, kaha joale re nkile qeto ea mokhoa, re ka koetlisa mohlala oa rona ka lintlha tsa koetliso.
# Train the decision tree classifier
model = DecisionTreeClassifier()
model.fit(X_train, y_train)
5. Hlahloba mohlala
Ka mor'a hore mohlala o koetlisoe, o tlameha ho hlahlojoa ka lintlha tse ncha ho netefatsa hore o nepahetse ebile oa tšepahala. Sena se kenyelletsa teko ea mohlala ka lintlha tse sa kang tsa sebelisoa nakong ea koetliso le ho bapisa litekanyetso tsa eona tse lekanyelitsoeng le boleng ba sebele boitsebisong ba tlhahlobo.
Tlhahlobo ena e ka thusa ho khetholla mefokolo leha e le efe ea mohlala, e kang ho feta kapa ho se sebetse hantle, 'me e ka lebisa ho lokisoeng leha e le hofe ho ka hlokahalang.
Re sebelisa lintlha tsa tlhahlobo, re tla lekola ho nepahala ha mohlala oa rona.
# 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)
Lintlha tse nepahetseng ha li mpe hakaalo hajoale. 🙂 Ho ntlafatsa lintlha tsa hau tsa ho nepahala, o ka lula o hloekisa lintlha haholoanyane kapa oa leka mefuta e fapaneng ea ho ithuta ka mochini ho bona hore na ke efe e fanang ka lintlha tse phahameng ka ho fetisisa.
6. Beha mohlala hantle
Haeba ts'ebetso ea mohlala e sa lekana, o ka e lokisa hantle ka ho fetola liparamente tse fapaneng tsa algorithm kapa ka ho etsa liteko ka li-algorithms tse ncha ka botlalo.
Ts'ebetso ena e ka kenyelletsa ho leka litekanyetso tse ling tsa ho ithuta, ho fetola maemo a tloaelehileng, kapa ho fetola palo kapa boholo ba likarolo tse patiloeng ho netweke ea methapo.
7. Sebelisa mohlala
Ha u se u khahliloe ke ts'ebetso ea mohlala, u ka qala ho e sebelisa ho hlahisa likhakanyo tsa data e ncha.
Sena se ka kenyelletsa ho fepa data e ncha mofuteng le ho sebelisa lintlha tse ithutoang tsa mohlala ho hlahisa likhakanyo ho data eo, kapa ho kopanya mohlala ho ts'ebeliso kapa sistimi e pharalletseng.
Re ka sebelisa mohlala oa rona ho etsa likhakanyo ho data e ncha ka mor'a hore re thabele ho nepahala ha eona. U ka leka litekanyetso tse fapaneng tsa bong le lilemo.
# Test the model with new data
new_data = [[25, 1], [30, 0]]
predictions = model.predict(new_data)
print("Predictions: ", predictions)
Phethela
Re qetile ho koetlisa mohlala oa rona oa pele oa ho ithuta mochini.
Ke tšepa hore u fumane e le molemo. Joale u ka leka ho sebelisa mefuta e fapaneng ea ho ithuta ea mochini joalo ka Linear Regression kapa Random Forest.
Ho na le li-dataset le mathata a mangata ho Kaggle haeba u batla ho ntlafatsa khouto ea hau le kutloisiso ea ho ithuta ka mochini.
Leave a Reply