def test_eg_five_uniform(self):
"""
Testing EG Five Uniform Industries
"""
myIndex = cm.Index()
# Number of observations
N = 50
# Number of areas
Na = 10
# Number of industries
Ni = 5
# uniform exposure
exposure = np.ones(N)
# single industry
industry = np.array(list(np.arange(0, Ni)) * Na)
# uniform area distribution
y_list = [[j for i in range(Ni)] for j in range(Na)]
area = np.array([y for x in y_list for y in x])
# create dataframe
d = {'Exposure': exposure, 'Area': area, 'Industry': industry}
data = pd.DataFrame(data=d)
results = myIndex.ellison_glaeser(data, Na, Ni)
print(self.shortDescription())
for i in range(Ni):
self.assertTrue(abs(results[i] - 0.0) < ERROR_MARGIN)
def test_eg_max_concentration(self):
"""
Testing EG Maximum Concentration
"""
myIndex = cm.Index()
# Number of observations
N = 400000
# Number of areas
Na = 200000
# Number of industries
Ni = 2
# uniform exposure
exposure = np.zeros(N)
exposure[0:199999] = 1
exposure[399999] = 1
exposure[399998] = 0.0001
# two industries
area = np.array(list(np.arange(0, Na)) * Ni)
# uniform area distribution
y_list = [[j for i in range(Na)] for j in range(Ni)]
industry = np.array([y for x in y_list for y in x])
# create dataframe
d = {'Exposure': exposure, 'Area': area, 'Industry': industry}
data = pd.DataFrame(data=d)
print(self.shortDescription())
self.assertTrue(
abs(myIndex.ellison_glaeser(data, Na, Ni)[0] - 0.0) < ERROR_MARGIN)
def test_ci_api(self):
"""
Testing that all implemented indexes have monotonic confidence intervals
"""
myIndex = cm.Index()
portfolio = np.random.normal(loc=10, scale=1, size=100)
methods = [['cr', 5], ['berger_parker'], ['hhi'], ['hk', 3],
['hoover'], ['gini'], ['shannon'], ['atkinson', 1.5],
['gei', 3], ['theil'], ['kolm', 2]]
print(self.shortDescription())
for method in methods:
if len(method) > 1:
lower_bound, value, upper_bound = myIndex.compute(
portfolio,
method[1],
index=method[0],
ci=0.95,
samples=1000)
else:
lower_bound, value, upper_bound = myIndex.compute(
portfolio, index=method[0], ci=0.95, samples=1000)
self.assertTrue(lower_bound <= value)
self.assertTrue(upper_bound >= value)
def test_atkinson_uniform(self):
myIndex = cm.Index()
vector = np.ones(1000000)
print(self.shortDescription())
self.assertTrue(
abs(myIndex.atkinson(vector, 0.5) - 0.0) < ERROR_MARGIN)
self.assertTrue(abs(myIndex.atkinson(vector, 1) - 0.0) < ERROR_MARGIN)
self.assertTrue(abs(myIndex.atkinson(vector, 2) - 0.0) < ERROR_MARGIN)
def test_hhi(self):
"""
Testing Herfindahl-Hirschman
"""
myIndex = cm.Index()
vector = np.ones(10)
print(self.shortDescription())
self.assertTrue(abs(myIndex.hhi(vector) - 0.0) < ERROR_MARGIN)
def test_gini(self):
"""
Testing Gini
"""
myIndex = cm.Index()
vector = np.ones(10)
print(self.shortDescription())
self.assertTrue(abs(myIndex.gini(vector) - 0.0) < ERROR_MARGIN)
def test_shannon(self):
"""
Testing Shannon
"""
myIndex = cl.Index()
vector = np.ones(10)
print(self.shortDescription())
self.assertTrue(abs(myIndex.shannon(vector) - 0.0) < ERROR_MARGIN)
def test_cr(self):
"""
Testing Concentration Ratio
"""
myIndex = cm.Index()
n = 1000
vector = np.ones(n)
print(self.shortDescription())
self.assertTrue(abs(myIndex.cr(vector, 1) - 1.0 / n) < ERROR_MARGIN)
def test_hk(self):
"""
Testing Hannah-Kay
"""
myIndex = cm.Index()
n = 10
vector = np.ones(n)
print(self.shortDescription())
self.assertTrue(abs(myIndex.hk(vector, 1) - 1.0 / n) < ERROR_MARGIN)
self.assertTrue(abs(myIndex.hk(vector, 3) - 1.0 / n) < ERROR_MARGIN)
def test_bp(self):
"""
Testing Berger-Parker
"""
myIndex = cm.Index()
n = 1000
vector = np.ones(n)
print(self.shortDescription())
self.assertTrue(
abs(myIndex.cr(vector, 1) -
myIndex.berger_parker(vector)) < ERROR_MARGIN)
def test_compare_atkinson_with_R(self):
""" Atkinson Index: Comparison with R version in ineq package
Results
Atkinson a=0.5: 0.1796591
Atkinson a=1: 0.3518251
"""
myIndex = cm.Index()
x = np.array(
[541, 1463, 2445, 3438, 4437, 5401, 6392, 8304, 11904, 22261])
ERROR_MARGIN = 1e-5
print(self.shortDescription())
self.assertTrue(
abs(myIndex.atkinson(x, 0.5) - 0.1796591) < ERROR_MARGIN)
self.assertTrue(
abs(myIndex.atkinson(x, 1.0) - 0.3518251) < ERROR_MARGIN)
def test_eg_single_industry(self):
"""
Testing EG Single Industry
"""
myIndex = cm.Index()
# Number of observations
N = 10
# Number of areas
Na = 10
# Number of industries
Ni = 1
# uniform exposure
exposure = np.ones(N)
# single industry
industry = np.zeros(N, dtype=np.int)
# uniform area distribution
area = np.arange(0, N)
# create dataframe
d = {'Exposure': exposure, 'Area': area, 'Industry': industry}
data = pd.DataFrame(data=d)
print(self.shortDescription())
self.assertTrue(
abs(myIndex.ellison_glaeser(data, Na, Ni)[0] - 0.0) < ERROR_MARGIN)
# encoding: utf-8 # (c) 2016-2020 Open Risk, all rights reserved # # ConcentrationMetrics is licensed under the MIT license a copy of which is included # in the source distribution of TransitionMatrix. This is notwithstanding any licenses of # third-party software included in this distribution. You may not use this file except in # compliance with the License. # # Unless required by applicable law or agreed to in writing, software distributed under # the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, # either express or implied. See the License for the specific language governing permissions and # limitations under the License. import concentrationMetrics as cm import numpy as np myIndex = cm.Index() x = np.array([541, 1463, 2445, 3438, 4437, 5401, 6392, 8304, 11904, 22261]) # Comparison with R version in ineq package # Results # Gini: 0.4620911 # Atkinson a=0.5: 0.1796591 # Atkinson a=1: 0.3518251 print(myIndex.gini(x)) print(myIndex.atkinson(x, 0.5)) print(myIndex.atkinson(x, 1.0))
| Shannon|
|--------:|
| 4.018412|
| 3.848471|
| 3.814059|
| 3.976563|
| 3.969940|
| 3.776575|
"""
BCI = pd.read_json(dataset_path + "BCI.json")
y = BCI.values
# Dataframe API
myGroupIndex = cm.Index(data=y, index='simpson')
myGroupIndex.print(6)
myGroupIndex = cm.Index(data=y, index='invsimpson')
myGroupIndex.print(6)
myGroupIndex = cm.Index(data=y, index='shannon')
myGroupIndex.print(6)
# Columnwise API
mySingleIndex = cm.Index()
for i in range(6):
print('HHI')
print('----------')
print(i, mySingleIndex.hhi(y[i, :]))
import concentrationMetrics as cl import numpy as np myIndex = cl.Index() x = np.array([541, 1463, 2445, 3438, 4437, 5401, 6392, 8304, 11904, 22261]) # Comparison with R version in ineq package # Results # Gini: 0.4620911 # Atkinson a=0.5: 0.1796591 # Atkinson a=1: 0.3518251 print(myIndex.gini(x)) print(myIndex.atkinson(x, 0.5)) print(myIndex.atkinson(x, 1.0))
def test_atkinson(self):
myIndex = cm.Index()
vector = np.zeros(1000)
vector[0] = 1
print(self.shortDescription())
self.assertTrue(abs(myIndex.atkinson(vector, 2) - 1.0) < ERROR_MARGIN)
