linear_regression_gd.py

# Importar las librerías 
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

# Importar el dataset de entranamiento
dataset = pd.read_csv('Admission_Predict_Ver1.1.csv')
X = dataset.iloc[:len(dataset), 1].values
X = X.reshape(-1,1)
X = np.insert(X, 0, 1, axis = 1)
y = dataset.iloc[:len(dataset), -1].values.reshape(-1,1)

# Seleccionar conjunto de training y test
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.25, 
                                                    random_state = 0)

# Escalado de las variables
from sklearn.preprocessing import StandardScaler
st_x = StandardScaler()
X_train = st_x.fit_transform(X_train)
X_test = st_x.transform(X_test)
st_y = StandardScaler()
y_train = st_y.fit_transform(y_train).reshape(-1)
y_test = st_y.transform(y_test).reshape(-1)

# Aplicación del modelo con grad
regression = LinearRegressionGD(l_rate = 0.01, n_iter = 20000)
coef = regression.fit(X_train, y_train, np.array([0.0, 1]))
y_predict = regression.predict(X_test)

# Aplicación del modelo con librería de sklearn
from sklearn.linear_model import LinearRegression
regression_py = LinearRegression() 
regression_py.fit(X_train, y_train)
y_predict_py = regression_py.predict(X_test) 

# Gráfica de regresión conjunto de test
plt.scatter(X_test[:,1], y_test, color = "red") 
plt.plot(X_test[:,1], y_predict_py, color = "blue")
plt.legend(('Descenso de Gradiente',),
            loc='lower right')
plt.title("Probabilidad admisión vs GRE Score (Conjunto de Test)")
plt.xlabel("GRE Score")
plt.ylabel("Probabilidad admisión a Postgrado")
plt.show()