머신러닝.LogisticRegression.C option 변화 관찰

https://broscoding.tistory.com/132

머신러닝.linear_model.LogisticRegression(Class=3)

C가 커질 수록 과적합 된다

C가 커질수록 세밀하게 나눠준다

1/C=a

c:(cost,lose,penalty)

for j in range(-5,5):
    from sklearn.linear_model import LogisticRegression
    model = LogisticRegression(C=1*(10**j))
    model.fit(X,y)
    import mglearn
    plt.figure(figsize=[10,8])
    mglearn.plots.plot_2d_classification(model,X,cm='Reds',alpha=0.3)
    mglearn.discrete_scatter(X[:,0],X[:,1],y,alpha=0.7)

    w=model.coef_
    b=model.intercept_

    rng=np.array([X[:,0].min(),X[:,0].max()])
    for i in range(3):
        plt.plot(rng,-(w[i,0]*rng+b[i])/w[i,1],':',lw=4,label='line'+str(i))
    plt.legend()
    plt.title('C='+str(1*(10**j)))