1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
|
from analytics.tranche_functions import BCloss_recov_dist, GHquad
import numpy as np
import pandas as pd
import matplotlib as plt
gridsize = 500
def run_scenario(pool_size, rho, successprob, issuerweights, amount):
#settings and running
Z, w = GHquad(gridsize)
S, _ = BCloss_recov_dist(successprob, issuerweights, amount, rho, Z, w, Ngrid=gridsize+1, defaultflag=False)
raised = np.arange(0, pool_size+pool_size/gridsize, step=pool_size/gridsize)
df = pd.DataFrame(S.flatten(), columns=['Probability'], index=raised)
return df
def plot_scenarios(df, pool_size):
''' takes run_scenario df '''
bins = [0,0.01,50,100,150,200,250,300,350,400,450,500,pool_size]
df1 = df.groupby(pd.cut(df.index,bins,right=False)).sum()
ax = df1.plot(kind='bar')
plt.pyplot.tight_layout()
return ax
def prob_over(df, amount):
return df[amount:].sum()
def plot_prob_over(df):
ax = (1-df.cumsum()).plot()
ax.set_title('Probability of raising X amount of AUM')
ax.set_xlabel('Amount Raised')
ax.set_ylabel('Probability')
return ax
def add_to_plot(df, ax):
ax.plot(1-df.cumsum())
return ax
|
