Add dynamic covariates PMF cpp implementation for pnbd model#284
Open
aramisfacchinetti wants to merge 2 commits intobachmannpatrick:feature-pnbddyncov-pmffrom
Open
Add dynamic covariates PMF cpp implementation for pnbd model#284aramisfacchinetti wants to merge 2 commits intobachmannpatrick:feature-pnbddyncov-pmffrom
aramisfacchinetti wants to merge 2 commits intobachmannpatrick:feature-pnbddyncov-pmffrom
Conversation
- Introduced a new header file `pnbd_dyncov_pmf.h` containing the declaration of functions and classes for handling dynamic covariates in the pnbd model. - Implemented the `DynamicCovariates` class to manage and compute cumulative sums of covariate data. - Added multiple PMF functions to calculate probabilities based on dynamic covariates, including per-customer calculations and hypergeometric functions. - Ensured compatibility with Armadillo and GSL libraries for efficient mathematical computations.
- Introduced minimal working example for PNBD Dynamic Covariates PMF in `minimal_pmf_example.R`. - Created a comprehensive benchmark script `pmf_example_benchmark.R` to compare R and Rcpp implementations of PMF. - Added quick test script `quick_test_pmf.R` to verify function availability and basic functionality. - Developed a simple demonstration script `simple_pmf_demo.R` showcasing PMF functions and their usage. - Updated documentation for internal functions related to PMF calculations, ensuring clarity on parameters and outputs. - Enhanced internal functions for better performance and stability in PMF calculations.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This pull request introduces significant enhancements to the Pareto/NBD model with dynamic covariates in the
CLVToolspackage. The changes include new functionality for calculating the probability mass function (PMF) with dynamic covariates, additional helper functions, C++ integration for computational efficiency, and improved data preparation for covariate-based modelling.New functionality for dynamic covariates:
Added a new function
pnbd_dyncov_pmf_per_customerto calculate the PMF for the Pareto/NBD model with dynamic covariates for a single customer. This function integrates dynamic transaction and lifetime covariates into the PMF computation. (R/RcppExports.R,src/RcppExports.cpp,src/pnbd_dyncov_pmf.h) [1] [2] [3]Introduced helper functions for calculating covariate effects (
Bbar,Dbar,S1,S2_ij, etc.) and time boundaries (bu_i) used in the PMF computation. These functions support detailed modelling of customer behaviour over time. (R/RcppExports.R,src/pnbd_dyncov_pmf.h) [1] [2]Data preparation and R integration:
Implemented
pnbd_dyncov_prepare_datato preprocess transaction and covariate data for PMF computation. This function handles merging, transformations, and computation of time intervals and covariate effects. (R/pnbd_dyncov_pmf.R)Added an R interface function
pnbd_dyncov_pmfwith an option to use either the R or C++ implementation for PMF calculation. This allows flexibility in balancing ease of use and computational performance. (R/pnbd_dyncov_pmf.R)C++ integration for performance:
Introduced a C++ implementation of the dynamic covariate PMF computation to improve performance for large datasets. This includes the
pnbd_dyncov_pmf_per_customerfunction and supporting components. (src/pnbd_dyncov_pmf.h,src/RcppExports.cpp) [1] [2] [3]Added a wrapper for the GSL hypergeometric function
2F1with error handling to support numerical stability in PMF calculations. (R/RcppExports.R,src/pnbd_dyncov_pmf.h) [1] [2]Refactoring and renaming:
pnbd_dyncov_pmftopnbd_dyncov_pmf_rto distinguish the R implementation from the new interface function. (R/pnbd_dyncov_pmf.R)