Contents

1 one-sample

2 two-sample

3 paired two-sample

• 귀무가설: 모수와 같다.
  
• 대립가설: 모수와 다르다.
  

import numpy as np
import scipy as sp

#random number
N = 100
mu = 100
sd = 10
np.random.seed(0)
x = np.random.normal(mu, sd, N)

#histogram
import seaborn as sns
sns.distplot(x, kde=False, fit=sp.stats.norm)
plt.show()

#t-test: 1-sample
sp.stats.ttest_1samp(x, popmean=100) #모수의 평균이 100이라는 것을 알고 있는 경우

#two-sample t-test: 분산이 같은 경우
import numpy as np
import scipy as sp

np.random.seed(0)
x1 = np.random.normal(100, 10, 100)
x2 = np.random.normal(97, 10, 100)

sns.distplot(x1, kde=False, fit=sp.stats.norm)
sns.distplot(x2, kde=False, fit=sp.stats.norm)
plt.show()

sp.stats.ttest_ind(x1, x2, equal_var=True)

#two-sample t-test: 분산이 다른 경우
import numpy as np
import scipy as sp

np.random.seed(0)
x1 = np.random.normal(100, 10, 100)
x2 = np.random.normal(97, 5, 100)

sns.distplot(x1, kde=False, fit=sp.stats.norm)
sns.distplot(x2, kde=False, fit=sp.stats.norm)
plt.show()

sp.stats.ttest_ind(x1, x2, equal_var=False)

#two-sample t-test: 대응표본
import numpy as np
import scipy as sp

np.random.seed(0)
x1 = np.random.normal(100, 10, 100)
x2 = np.random.normal(97, 10, 100)

sns.distplot(x1, kde=False, fit=sp.stats.norm)
sns.distplot(x2, kde=False, fit=sp.stats.norm)
plt.show()

sp.stats.ttest_rel(x1, x2)