1 OLS vs Ridge vs Lasso

2 데이터

3 OLS(ordinary least squares)

4 Ridge

5 Lasso

#iris 데이터세트 만들기
import numpy as np
import pandas as pd
from sklearn.datasets import load_iris
iris = load_iris()
iris.data
iris.feature_names
iris.target
iris.target_names


iris_df = pd.DataFrame(iris.data, columns=iris.feature_names)
iris_df["target"] = iris.target 
iris_df["target_names"] = iris.target_names[iris.target]
iris_df[:5]



#훈련세트, 테스트세트 나누기
from sklearn.model_selection import train_test_split
train_set, test_set = train_test_split(iris_df, test_size = 0.3)

train_set.shape
test_set.shape

#선형 회귀(최소제곱)
from sklearn.linear_model import LinearRegression as lm
model_ols = lm().fit(X=train_set.ix[:, [2]], y=train_set.ix[:, [3]])
print(model_ols.coef_)
print(model_ols.intercept_)


#plot
import matplotlib.pyplot as plt
plt.scatter(train_set.ix[:, [2]], train_set.ix[:, [3]],  color='black')
plt.plot(test_set.ix[:, [2]], model_ols.predict(test_set.ix[:, [2]]))

#Ridge: alpha값을 조정하여 과대/과소적합을 피한다.
from sklearn.linear_model import Ridge
model_ridge = Ridge(alpha=10).fit(X=train_set.ix[:, [2]], y=train_set.ix[:, [3]])

#점수
print(model_ridge.score(X=train_set.ix[:, [2]], y=train_set.ix[:, [3]]))
print(model_ridge.score(X=test_set.ix[:, [2]], y=test_set.ix[:, [3]]))

#plot
import matplotlib.pyplot as plt
plt.scatter(train_set.ix[:, [2]], train_set.ix[:, [3]],  color='black')
plt.plot(test_set.ix[:, [2]], model_ridge.predict(test_set.ix[:, [2]]))

#Lasso: alpha값을 조정하여 과대/과소적합을 피한다.
from sklearn.linear_model import Lasso
model_lasso = Lasso(alpha=0.1, max_iter=1000).fit(X=train_set.ix[:, [0,1,2]], y=train_set.ix[:, [3]])

#점수
print(model_lasso.score(X=train_set.ix[:, [0,1,2]], y=train_set.ix[:, [3]]))
print(model_lasso.score(X=test_set.ix[:, [0,1,2]], y=test_set.ix[:, [3]]))

#사용한 특성수
print(np.sum(model_lasso.coef_ != 0))