package useful for likelihood-based inference

hawkes-process likelihood profiling simulation-modeling statistics
4 Open Issues Need Help Last updated: Jun 22, 2025

Open Issues Need Help

View All on GitHub
Statistical Computing Likelihood Inference

AI Summary: Enhance the 'pykelihood' package to handle n-dimensional covariates in kernel functions. This involves addressing broadcasting issues for multi-dimensional inputs and potentially allowing users to specify covariates using named dataframes (similar to R's syntax). The solution needs to consider how SciPy handles n-dimensional parameters and ensure compatibility.

Complexity: 4/5
enhancement help wanted priority:high

package useful for likelihood-based inference

Python
#hawkes-process#likelihood#profiling#simulation-modeling#statistics
Statistical Computing Likelihood Inference

AI Summary: Implement a feature in the `pykelihood` Python package to handle Bayesian inference when a distribution is a parameter of another distribution. This involves automatically generating a PyMC model and performing MCMC optimization.

Complexity: 5/5
enhancement help wanted priority:low

package useful for likelihood-based inference

Python
#hawkes-process#likelihood#profiling#simulation-modeling#statistics
Statistical Computing Likelihood Inference

AI Summary: Implement support for sparse parameters within the pykelihood package. This involves enabling the use of sparse arrays for parameters and providing mechanisms to optimize only specific elements within multi-dimensional parameters, potentially through a flexible Parameter-protocol or case-by-case handling.

Complexity: 4/5
enhancement help wanted priority:low

package useful for likelihood-based inference

Python
#hawkes-process#likelihood#profiling#simulation-modeling#statistics
Statistical Computing Likelihood Inference

AI Summary: Implement parameter domain constraints for improved robustness in the pykelihood package. This involves adding metadata for parameter bounds, potentially using constraint-aware optimization algorithms (like COBYQA), or incorporating penalties for out-of-domain values, especially handling complex, multi-dimensional parameter relationships.

Complexity: 4/5
enhancement help wanted priority:medium

package useful for likelihood-based inference

Python
#hawkes-process#likelihood#profiling#simulation-modeling#statistics