Generating 100K model points¶

This notebook generates the sample model points ued by the “3. Memory-Optimized Mutiprocess Example” notebook. A DataFrame of 100,000 model points is generated and saved to an Excel file named model_point_table_100K.xlsx in the current directory. The generated model point file can be loaded back by calling pandas’ read_excel function with index_col=0. The DataFrame can be used with CashValue_ME by setting it to model_point_table in the Projection space.

Index and Columns:

point_id(Index): Model point identifier
spec_id: Character identifier (‘A’,‘B’,‘C’ and ‘D’) representing the product specs of the model point. product_spec_table in CashValue_ME.Projection defines the product specs for each identifier.
age_at_entry: Issue age. The samples are distributed uniformly from 20 to 59.
sex: “M” or “F” to indicate policy holder’s sex.
policy_term: Policy term in years. The samples are evenly distriubted among 10, 15 and 20.
policy_count: The number of policies. Uniformly distributed from 0 to 100.
sum_assured: Sum assured. The samples are uniformly distributed from 10,000 to 1,000,000.
duration_mth: Months elapsed from the issue til t=0. New business is assumed so 0 for all model points.
premium_pp: Premium per policy.
av_pp_init: Initial account value per policy for existing policies. 0 for all model points.

Number of model points:

100,000

[5]:

import numpy as np
from numpy.random import default_rng  # Requires NumPy 1.17 or newer

rng = default_rng(12345)

# Number of Model Points
MPCount = 100000

# Issue Age (Integer): 20 - 59 year old
age_at_entry = rng.integers(low=20, high=60, size=MPCount)

# Sex (Char)
Sex = ["M", "F"]

sex = np.fromiter(map(lambda i: Sex[i], rng.integers(low=0, high=len(Sex), size=MPCount)), np.dtype('<U1'))

# Policy Term (Integer, modified later for whole life): 10, 15, 20
policy_term = rng.integers(low=0, high=3, size=MPCount) * 5 + 10


# Sum Assured (Float): 10,000 - 1,000,000
sum_assured = np.round((1_000_000 - 10_000) * rng.random(size=MPCount) + 10_000, -3)


# Policy Count (Integer): 1
policy_count = np.rint(100 * rng.random(size=MPCount)).astype(int)

[6]:

import pandas as pd

attrs = [
    "age_at_entry",
    "sex",
    "policy_term",
    "policy_count",
    "sum_assured"
]

data = [
    age_at_entry,
    sex,
    policy_term,
    policy_count,
    sum_assured
]

model_point_table = pd.DataFrame(dict(zip(attrs, data)), index=range(1, MPCount+1))
model_point_table.index.name = "policy_id"

[7]:

# Spec ID
SpedIDs = "ABCD"

spec_ids = np.fromiter(map(lambda i: SpedIDs[i], rng.integers(low=0, high=len(SpedIDs), size=MPCount)), np.dtype('<U1'))
model_point_table.insert(0, 'spec_id', spec_ids)

whole_life = (model_point_table['spec_id'] == 'C')  | (model_point_table['spec_id'] == 'D')

# Modify Policy Term

model_point_table['policy_term'].mask(whole_life, 9999, inplace=True)

# Set Duration to 0

model_point_table['duration_mth'] = 0

# Premium per Plicy (by adjusting size between single premium(A and B) and level premium(C and D) policies)

premium_pp = np.ceil(model_point_table['sum_assured'] / (80 - model_point_table['age_at_entry']) / 12 / 100) * 100

model_point_table.insert(len(model_point_table.columns), 'premium_pp', model_point_table['sum_assured'])
model_point_table['premium_pp'].mask(whole_life, premium_pp, inplace=True)

# Ininital Account Value per Policy

model_point_table['av_pp_init'] = 0

[8]:

model_point_table

[8]:

	spec_id	age_at_entry	sex	policy_term	policy_count	sum_assured	duration_mth	premium_pp	av_pp_init
policy_id
1	A	47	M	20	22	804000.0	0	804000.0	0
2	C	29	F	9999	75	519000.0	0	900.0	0
3	A	51	F	10	5	409000.0	0	409000.0	0
4	B	32	M	15	60	128000.0	0	128000.0	0
5	D	28	M	9999	45	698000.0	0	1200.0	0
...	...	...	...	...	...	...	...	...	...
99996	A	21	M	10	34	152000.0	0	152000.0	0
99997	D	24	F	9999	53	928000.0	0	1400.0	0
99998	B	46	F	15	72	662000.0	0	662000.0	0
99999	A	46	M	15	36	583000.0	0	583000.0	0
100000	B	35	M	15	3	638000.0	0	638000.0	0

100000 rows × 9 columns

[9]:

model_point_table.to_excel('model_point_table_100K.xlsx')

3. Memory-Optimized Mutiprocess Example

The assets Library