# Ratemaking
A comprehensive Python library for Property & Casualty actuarial ratemaking, providing tools for credibility analysis, trending, exposure calculations, and data processing.
[](https://www.python.org/downloads/)
[](https://opensource.org/licenses/MIT)
## Features
### Currently Available
- **Credibility Analysis**: Classical, Bühlmann, and Bayesian credibility methods
- **Complement Calculations**: First-Dollar methods from Werner & Modlin Chapter 12
- **Comprehensive Testing**: Test suite with actuarial validation
## Installation
```bash
pip install ratemaking
```
## Quick Start
### Classical Credibility
```python
from ratemaking import (
classical_full_credibility_frequency,
classical_partial_credibility
)
# Calculate full credibility standard
n_full = classical_full_credibility_frequency(p=0.95, k=0.05)
# Calculate credibility factor
z = classical_partial_credibility(n=observed_claims, n_full=n_full)
# Apply credibility blend
estimate = z * observed_rate + (1 - z) * complement_rate
```
### Bühlmann Credibility
```python
from ratemaking import BuhlmannInputs, buhlmann
data = {"risk_1": [1.2, 1.5], "risk_2": [2.1, 1.9]}
result = buhlmann(BuhlmannInputs(data=data))
print(f"Credibility weights: {result.Z_by_risk}")
```
### Bayesian Credibility
```python
from ratemaking import bayes_poisson_gamma
# Poisson-Gamma conjugate updating
posterior = bayes_poisson_gamma(
prior_alpha=2.0, prior_beta=100.0,
total_counts=15, total_exposure=120
)
print(f"Posterior mean: {posterior.mean}")
print(f"Credibility weight: {posterior.credibility_Z}")
```
### Complement Calculations
```python
from ratemaking.complements import (
trended_present_rates_loss_cost,
trended_present_rates_rate_change_factor,
larger_group_applied_rate_change_to_present_rate,
HarwayneInputs,
harwayne_complement,
)
# Trended present rates method
complement = trended_present_rates_loss_cost(
present_rate=100.0,
prior_indicated_factor=1.10, # 10% indicated
prior_implemented_factor=1.06, # 6% implemented
loss_trend_annual=0.05, # 5% annual loss trend
trend_years=2.0
)
# Harwayne's method for multi-state complements
inputs = HarwayneInputs(
target_class_exposures=target_exposures,
target_avg_pure_premium=120.0,
related_state_class_pp=related_pp_data,
related_state_class_exposures=related_exposure_data,
class_of_interest='ClassA'
)
complement = harwayne_complement(inputs)
```
## Package Structure
```
ratemaking/
├── credibility/ # Credibility analysis tools
│ ├── classical.py # Classical (Limited Fluctuation) credibility
│ ├── buhlmann.py # Bühlmann & Bühlmann-Straub credibility
│ └── bayesian.py # Bayesian credibility with conjugate priors
├── complements/ # Complement calculation methods
│ └── first_dollar.py # First-Dollar methods (Werner & Modlin Ch.12)
├── trending/ # Trending analysis tools (coming soon)
├── exposure/ # Exposure calculation tools (coming soon)
└── utils/ # Data processing utilities (coming soon)
```
## Modular Usage
For organized imports, use the submodules:
```python
# Organized by functionality
from ratemaking.credibility import classical, buhlmann, bayesian
from ratemaking.complements import first_dollar
# Use specific functions
n_full = classical.classical_full_credibility_frequency(p=0.95, k=0.05)
complement = first_dollar.trended_present_rates_loss_cost(...)
```
## Testing
Run the test suite:
```bash
pytest tests/ -v
```
## Development
### Setting up for development:
```bash
git clone https://github.com/little-croissant/ratemaking.git
cd ratemaking
pip install -e .
pip install -e ".[test]"
```
## License
MIT License
Raw data
{
"_id": null,
"home_page": "https://github.com/little-croissant/ratemaking",
"name": "ratemaking",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.8",
"maintainer_email": null,
"keywords": "actuarial ratemaking credibility complements trending insurance P&C casualty property",
"author": "Hugo Latendresse",
"author_email": "hugolatendresse@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/ae/3a/0f3c0fb52d5c561a1200a040c4e4efe98265fd0f281d5d97f0b92c7ac0fc/ratemaking-0.3.0.tar.gz",
"platform": null,
"description": "# Ratemaking\n\nA comprehensive Python library for Property & Casualty actuarial ratemaking, providing tools for credibility analysis, trending, exposure calculations, and data processing.\n\n[](https://www.python.org/downloads/)\n[](https://opensource.org/licenses/MIT)\n\n## Features\n\n### Currently Available\n- **Credibility Analysis**: Classical, B\u00fchlmann, and Bayesian credibility methods\n- **Complement Calculations**: First-Dollar methods from Werner & Modlin Chapter 12\n- **Comprehensive Testing**: Test suite with actuarial validation\n\n## Installation\n\n```bash\npip install ratemaking\n```\n\n## Quick Start\n\n### Classical Credibility\n\n```python\nfrom ratemaking import (\n classical_full_credibility_frequency,\n classical_partial_credibility\n)\n\n# Calculate full credibility standard\nn_full = classical_full_credibility_frequency(p=0.95, k=0.05)\n\n# Calculate credibility factor \nz = classical_partial_credibility(n=observed_claims, n_full=n_full)\n\n# Apply credibility blend\nestimate = z * observed_rate + (1 - z) * complement_rate\n```\n\n### B\u00fchlmann Credibility\n\n```python\nfrom ratemaking import BuhlmannInputs, buhlmann\n\ndata = {\"risk_1\": [1.2, 1.5], \"risk_2\": [2.1, 1.9]}\nresult = buhlmann(BuhlmannInputs(data=data))\nprint(f\"Credibility weights: {result.Z_by_risk}\")\n```\n\n### Bayesian Credibility\n\n```python\nfrom ratemaking import bayes_poisson_gamma\n\n# Poisson-Gamma conjugate updating\nposterior = bayes_poisson_gamma(\n prior_alpha=2.0, prior_beta=100.0,\n total_counts=15, total_exposure=120\n)\nprint(f\"Posterior mean: {posterior.mean}\")\nprint(f\"Credibility weight: {posterior.credibility_Z}\")\n```\n\n### Complement Calculations\n\n```python\nfrom ratemaking.complements import (\n trended_present_rates_loss_cost,\n trended_present_rates_rate_change_factor,\n larger_group_applied_rate_change_to_present_rate,\n HarwayneInputs,\n harwayne_complement,\n)\n\n# Trended present rates method\ncomplement = trended_present_rates_loss_cost(\n present_rate=100.0,\n prior_indicated_factor=1.10, # 10% indicated\n prior_implemented_factor=1.06, # 6% implemented \n loss_trend_annual=0.05, # 5% annual loss trend\n trend_years=2.0\n)\n\n# Harwayne's method for multi-state complements\ninputs = HarwayneInputs(\n target_class_exposures=target_exposures,\n target_avg_pure_premium=120.0,\n related_state_class_pp=related_pp_data,\n related_state_class_exposures=related_exposure_data,\n class_of_interest='ClassA'\n)\ncomplement = harwayne_complement(inputs)\n```\n\n## Package Structure\n\n```\nratemaking/\n\u251c\u2500\u2500 credibility/ # Credibility analysis tools\n\u2502 \u251c\u2500\u2500 classical.py # Classical (Limited Fluctuation) credibility\n\u2502 \u251c\u2500\u2500 buhlmann.py # B\u00fchlmann & B\u00fchlmann-Straub credibility\n\u2502 \u2514\u2500\u2500 bayesian.py # Bayesian credibility with conjugate priors\n\u251c\u2500\u2500 complements/ # Complement calculation methods\n\u2502 \u2514\u2500\u2500 first_dollar.py # First-Dollar methods (Werner & Modlin Ch.12)\n\u251c\u2500\u2500 trending/ # Trending analysis tools (coming soon)\n\u251c\u2500\u2500 exposure/ # Exposure calculation tools (coming soon)\n\u2514\u2500\u2500 utils/ # Data processing utilities (coming soon)\n```\n\n## Modular Usage\n\nFor organized imports, use the submodules:\n\n```python\n# Organized by functionality\nfrom ratemaking.credibility import classical, buhlmann, bayesian\nfrom ratemaking.complements import first_dollar\n\n# Use specific functions\nn_full = classical.classical_full_credibility_frequency(p=0.95, k=0.05)\ncomplement = first_dollar.trended_present_rates_loss_cost(...)\n```\n\n## Testing\n\nRun the test suite:\n\n```bash\npytest tests/ -v\n```\n\n## Development\n\n### Setting up for development:\n\n```bash\ngit clone https://github.com/little-croissant/ratemaking.git\ncd ratemaking\npip install -e .\npip install -e \".[test]\"\n```\n\n## License\n\nMIT License \n",
"bugtrack_url": null,
"license": null,
"summary": "A comprehensive Python library for P&C actuarial ratemaking",
"version": "0.3.0",
"project_urls": {
"Bug Reports": "https://github.com/little-croissant/ratemaking/issues",
"Documentation": "https://github.com/little-croissant/ratemaking#readme",
"Homepage": "https://github.com/little-croissant/ratemaking",
"Source": "https://github.com/little-croissant/ratemaking"
},
"split_keywords": [
"actuarial",
"ratemaking",
"credibility",
"complements",
"trending",
"insurance",
"p&c",
"casualty",
"property"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "eed1e49bd2c80953e72db3f4f11eaf847eaa66f3d620d78986b700963ab7b9ee",
"md5": "280cb8b23d771322fcf618362b14cc4d",
"sha256": "91f8bdacd33c2aa4d006c88ac72734ee995543125cf59463852808f655b18964"
},
"downloads": -1,
"filename": "ratemaking-0.3.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "280cb8b23d771322fcf618362b14cc4d",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.8",
"size": 22003,
"upload_time": "2025-10-26T21:01:58",
"upload_time_iso_8601": "2025-10-26T21:01:58.194908Z",
"url": "https://files.pythonhosted.org/packages/ee/d1/e49bd2c80953e72db3f4f11eaf847eaa66f3d620d78986b700963ab7b9ee/ratemaking-0.3.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "ae3a0f3c0fb52d5c561a1200a040c4e4efe98265fd0f281d5d97f0b92c7ac0fc",
"md5": "83107e2b1fa28e031556c7e6e5f7f3a6",
"sha256": "b0b4ec0c2ce4d37c4e73e7bf11fdc6a6a6b46f767fd370ae22d6bbaf42d1be54"
},
"downloads": -1,
"filename": "ratemaking-0.3.0.tar.gz",
"has_sig": false,
"md5_digest": "83107e2b1fa28e031556c7e6e5f7f3a6",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.8",
"size": 18972,
"upload_time": "2025-10-26T21:01:59",
"upload_time_iso_8601": "2025-10-26T21:01:59.193256Z",
"url": "https://files.pythonhosted.org/packages/ae/3a/0f3c0fb52d5c561a1200a040c4e4efe98265fd0f281d5d97f0b92c7ac0fc/ratemaking-0.3.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-10-26 21:01:59",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "little-croissant",
"github_project": "ratemaking",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"requirements": [
{
"name": "pytest",
"specs": [
[
">=",
"7.0.0"
]
]
},
{
"name": "numpy",
"specs": [
[
">=",
"1.20.0"
]
]
},
{
"name": "pandas",
"specs": [
[
">=",
"1.3.0"
]
]
},
{
"name": "pyperclip",
"specs": [
[
">=",
"1.8.0"
]
]
},
{
"name": "pyautogui",
"specs": [
[
">=",
"0.9.0"
]
]
},
{
"name": "watchdog",
"specs": [
[
">=",
"2.0.0"
]
]
}
],
"lcname": "ratemaking"
}
JSON
