LPCDE

The lpcde Python package implements local polynomial conditional distribution and density estimation with bandwidth selection.

• lpcde: local polynomial conditional CDF, PDF, and derivative estimation with pointwise and uniform inference quantities.
• lpbwcde: rule-of-thumb bandwidth selection for local polynomial conditional density estimation.

Website: https://nppackages.github.io/lpcde/.

Source code: https://github.com/nppackages/lpcde.

Authors

Matias D. Cattaneo (matias.d.cattaneo@gmail.com)

Rajita Chandak (rajita.chandak@gmail.com)

Michael Jansson (michael.jansson.berkeley@gmail.com)

Xinwei Ma (xinweima.pku@gmail.com)

Installation

To install locally from the repository:

pip install -e Python/lpcde

Usage

import numpy as np
from lpcde import confint, lpbwcde, lpcde, plot, summary, vcov

x_data = np.random.normal(size=500)
y_data = np.random.normal(loc=x_data, scale=1)
y_grid = np.linspace(-1, 1, 5)

fit = lpcde(x_data=x_data, y_data=y_data, y_grid=y_grid, x=0, bw=0.5)
bw = lpbwcde(y_data=y_data, x_data=x_data, x=0, y_grid=y_grid)
summary(fit)
confint(fit)

Replication examples are available in replication/lpcde_illustration.py and replication/lpcde_replication.py.

Dependencies

• numpy
• pandas
• scipy
• matplotlib

References

Software and Implementation

• Cattaneo, Chandak, Jansson and Ma (2025): lpcde: Estimation and Inference for Local Polynomial Conditional Density Estimators.
  Journal of Open Source Software 10(107): 7241.

Technical and Methodological

• Cattaneo, Chandak, Jansson and Ma (2024): Boundary Adaptive Local Polynomial Conditional Density Estimators.
  Bernoulli 30(4): 3193-3223.