linear-discriminant-analysis-with-python-scikit-learn/linear_discriminant_analysis.py

# conda activate PANDA

# https://wellsr.com/python/linear-discriminant-analysis-for-dimensionality-reduction-in-python/

import pandas as pd
import numpy as np
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score
from sklearn.discriminant_analysis import LinearDiscriminantAnalysis as LDA
from sklearn.decomposition import PCA


header_list = ["Preg", "Glucose", "BP", "skinThick", "Insulin", "BMI", "DPF", "Age", "Class"]
diabetes_ds = pd.read_csv(r"pima-indians-diabetes.csv", names = header_list)

print(diabetes_ds.head())

features = diabetes_ds.drop(['Class'], axis=1)
labels = diabetes_ds["Class"]

X_train, X_test, y_train, y_test = train_test_split(features, labels,  test_size=0.20, random_state=0)

# print(X_train.head(6))
X_Test = X_test

stand_scal = StandardScaler()
X_train = stand_scal.fit_transform(X_train)
X_test = stand_scal.transform(X_test)

model = RandomForestClassifier(n_estimators=50, random_state=0)
model.fit(X_train, y_train)
y_pred = model.predict(X_test)

# print('accuracy Default Features = ', accuracy_score(y_test, y_pred))
# print("X_test[13] = ", X_test[13])
# print('y_test = ', y_test.head(13))
# print(y_pred)



# Dimensionality Reduction with Linear Discriminant Analysis

lda_model = LDA(n_components = 1)
X_train_lda = lda_model.fit_transform(X_train, y_train)
X_test_lda = lda_model.transform(X_test)

# print(X_train_lda.shape)
# print(X_test_lda.shape)

model = RandomForestClassifier(n_estimators=50, random_state=0)
model.fit(X_train_lda, y_train)
y_pred = model.predict(X_test_lda)

# print('accuracy LDA = ', accuracy_score(y_test, y_pred))

# print("X_test_lda[13] = ", X_test_lda[13])
# # print(y_test.head(13))
# print('y_pred = ', y_pred)

# print('y_pred[13] = ', y_pred[13])



# Comparing LDA to PCA

pca_model = PCA(n_components= 1)
X_train_pca = pca_model.fit_transform(X_train)
X_test_pca = pca_model.transform(X_test)

model = RandomForestClassifier(n_estimators=50, random_state=0)
model.fit(X_train_pca, y_train)
y_pred = model.predict(X_test_pca)

# print('accuracy PCA = ', accuracy_score(y_test, y_pred))
# print("X_train_pca[13] = ", X_train_pca[13])
# print('y_pred = ', y_pred)

# print('y_pred[13] = ', y_pred[13])

'''
Приведенный выше результат показывает, что при сокращении набора признаков до 1 компонента
с помощью метода анализа главных компонент (PCA) мы получаем точность 66,23 %, что примерно
на 11 % меньше, чем 77,92 %, достигнутых с помощью метода LDA. Это ясно показывает, что для
нашего набора данных LDA является более подходящей методикой сокращения размерности, и
должно убедить вас в том, что вам нужно играть с различными методиками при сокращении
размерности в ваших собственных наборах данных.
'''

print('y_pred = ', y_pred)
print('X_Test = \n', X_Test.head(6))
X_Test['pred'] = y_pred
# X_Test.to_csv('data1.csv')
print('X_Test = \n', X_Test.head(6))