Isingeniso
I-Artificial Neural Networks (ANN) ilethe uguquko olukhulu emkhakheni wokufunda komshini. Izakhiwo zenethiwekhi ezifaniswe ebuchosheni bomuntu, ezisetshenziselwa ukuxazulula imisebenzi eyinkimbinkimbi efana nokubonwa kwesithombe, ukuhumusha inkulumo, nokuningi. Amandla ama-ANN afinyeleleke ngelabhulali yePython, Keras. I-Keras Sequential Model kulula kakhulu ukuyiqonda nokusebenzisa, ihlinzeka ngamakhodi afundekayo ngabantu. Le ndatshana izocubungula i-Keras Sequential Model futhi iveze ukuthi ungasebenzisa kanjani imitapo yolwazi ku-ANN.
Ukhiye Wokuvula ama-ANN: Imodeli Elandelanayo ye-Keras Nemitapo yolwazi
I-Keras Sequential Model iyinqwaba yezendlalelo ongazisebenzisa ukuze udale inethiwekhi ye-neural. Ilula futhi ilungele abaqalayo ngoba ikuvumela ukuthi wakhe imodeli ngesinyathelo ngesinyathelo. Nokho, ayilungele izakhiwo eziyinkimbinkimbi. Isibonelo, ayisekeli okokufaka okuningi noma izendlalelo ezabiwe nokunye.
Nge-Keras, ungaqala ngokwakha imodeli yokulandelana engenalutho bese wengeza izendlalelo kuyo, noma ungadlulisa uhlu lwezendlalelo lapho udala. Amalabhulali ayeza ukuze adlale uma ufuna ukwenza imisebenzi ethile efana nokukhohlisa kwe-matrix noma ukubunjwa kabusha kwedatha, phakathi kokunye.
from keras.models import Sequential from keras.layers import Dense # Initializing the ANN model = Sequential() # Adding the input layer and the first hidden layer model.add(Dense(16, activation='relu', input_shape=(10,))) # Adding the second hidden layer model.add(Dense(8, activation='relu')) # Adding the output layer model.add(Dense(1, activation='sigmoid'))
Lesi isibonelo esilula sokuthi ungaqala kanjani ngemodeli yokulandelana engenalutho bese wengeza izendlalelo kuyo.
Ukufaneleka Kwemitapo yolwazi
Amalabhulali yilapho amandla eqiniso olimi lokuhlela lwe-Python ebonakala khona. Ngemitapo yolwazi efana ne-NumPy nama-panda, ukuphathwa kwedatha kanye nemisebenzi yezibalo kuyathuthuka futhi kube lula ukuyiphatha. Amanye amalabhulali afana ne-Scikit-learn ahambisana ne-Keras ekwakheni i-ANN ngokuhlukanisa amasethi edatha kanye nokunikeza amamethrikhi.
import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
# Data Preparation
data = pd.read_csv('my_data.csv')
X = data.iloc[:, :-1].values
y = data.iloc[:, -1].values
# Splitting the dataset into Training set and Test set
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 0)
Imitapo yolwazi esetshenziswe ngokuphawulekayo yenza kube lula inqubo yokulungiselela idatha edingekayo ekwakheni i-ANN.
Ivumela Ukucutshungulwa Kwedatha nge-TensorFlow
Ku-Python, i-TensorFlow yakha isisekelo sokusebenza namaKeras. Iwumtapo wolwazi obalulekile wokudala nokuqeqesha i-ANN njengoba inikeza uhlaka lokuqalisa amanethiwekhi e-neural asezingeni eliphezulu. Umsebenzi we-TensorFlow ukuhlinzeka ngomthamo wokubala ngezinombolo kusetshenziswa amagrafu agelezayo wedatha anama-node nemiphetho.
Ngomongo we-Keras Sequential Model, i-TensorFlow isiza ekubaleni ukusebenza kwezibalo okuyinkimbinkimbi okwenzeka ngemuva kwemodeli.
# Compiling the ANN model.compile(optimizer ='adam',loss='binary_crossentropy', metrics =['accuracy']) # Fitting the ANN to the Training set model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=100, batch_size=32)
Lokhu, empeleni, kumelela indlela i-TensorFlow ebaluleke ngayo enqubweni yokufunda yemodeli elandelanayo kuma-Keras. Indima yayo ibalulekile ezigabeni zokuthuthukisa kanye nezibalo.
Ukuvela kwe-ANN okunikwa amandla yi-Python nemitapo yolwazi yayo ngemodeli ye-Keras Sequential ibonisa isigaba esijabulisayo sokufunda komshini. Ukukhethwa kwemitapo yolwazi nokuqonda ukusebenza kwayo kungaba nomthelela omkhulu empumelelweni yemodeli. Ngokuqhubeka nokuzijwayeza nokusebenzisa, ubuciko bokwenza amamodeli e-ANN asebenzayo kusetshenziswa i-Keras kuba imvelo yesibili kubahleli bezinhlelo ze-python.
