|
1
|
|
|
import math |
|
2
|
|
|
from enum import Enum |
|
3
|
|
|
|
|
4
|
|
|
from scipy.stats import norm |
|
5
|
|
|
|
|
6
|
|
|
|
|
7
|
|
|
class DistributionFamily(Enum): |
|
8
|
|
|
normal = 1 |
|
9
|
|
|
student_t = 2 |
|
10
|
|
|
fisher = 3 |
|
11
|
|
|
chi_square = 4 |
|
12
|
|
|
|
|
13
|
|
|
class MeanSamplingDistribution(object): |
|
14
|
|
|
sample_distribution = None |
|
15
|
|
|
point_estimate = None |
|
16
|
|
|
distribution_family = None |
|
17
|
|
|
df = None |
|
18
|
|
|
|
|
19
|
|
|
def __init__(self, sample_distribution=None, sample_mean=None, sample_sd=None, sample_size=None): |
|
20
|
|
|
if sample_mean is not None: |
|
21
|
|
|
self.point_estimate = sample_mean |
|
22
|
|
|
|
|
23
|
|
|
if sample_sd is not None: |
|
24
|
|
|
self.sample_sd = sample_sd |
|
25
|
|
|
|
|
26
|
|
|
if sample_size is not None: |
|
27
|
|
|
self.sample_size = sample_size |
|
28
|
|
|
|
|
29
|
|
|
if sample_distribution is not None: |
|
30
|
|
|
self.sample_distribution = sample_distribution |
|
31
|
|
|
self.point_estimate = sample_distribution.mean |
|
32
|
|
|
self.sample_sd = sample_distribution.sd |
|
33
|
|
|
self.sample_size = sample_distribution.sample_size |
|
34
|
|
|
|
|
35
|
|
|
self.standard_error = MeanSamplingDistribution.calculate_standard_error(self.sample_sd, self.sample_size) |
|
36
|
|
|
|
|
37
|
|
|
self.df = self.sample_size - 1.0 |
|
38
|
|
|
if self.sample_size < 30: |
|
39
|
|
|
self.distribution_family = DistributionFamily.student_t |
|
40
|
|
|
else: |
|
41
|
|
|
self.distribution_family = DistributionFamily.normal |
|
42
|
|
|
|
|
43
|
|
|
|
|
44
|
|
|
@staticmethod |
|
45
|
|
|
def calculate_standard_error(sample_sd, sample_size): |
|
46
|
|
|
return sample_sd / math.sqrt(sample_size) |
|
47
|
|
|
|
|
48
|
|
|
def confidence_interval(self, confidence_level): |
|
49
|
|
|
pf = norm.ppf(1 - (1 - confidence_level) / 2) |
|
50
|
|
|
return (-pf, pf) |
|
51
|
|
|
|
|
52
|
|
|
|
|
53
|
|
|
|
chen0040 /
pysie
