text summarizing the content or functions

Author: vlyubchich

R/VAR.R

@@ -1,10 +1,20 @@
-# This is a copy of the code from the package vars by Bernhard Pfaff
-# with changes that were not merged to the package
-# with the pull request from 2021-08-22
+# Internal VAR Estimation Function (Modified from vars package)
+#
+# This is a modified copy of the VAR() function from the vars package by Bernhard Pfaff.
+# The modifications were submitted via pull request on 2021-08-22 but were not merged:
 # https://github.com/bpfaff/vars/pull/10/files
-# I.e., subsequent vars v.1.5-6 (2021-09-17) does not include these changes.
-# The changes (most of them are marked with "#VL") allow for restricted VAR
-# estimation and Granger causality testing using such models.
+#
+# The vars package v.1.5-6 (2021-09-17) and later versions do not include these changes.
+# This modified version is maintained internally to support restricted VAR estimation
+# and Granger causality testing with lag restrictions.
+#
+# Key Modifications (marked with "#VL" comments):
+# - Added lag.restrict parameter to exclude near-contemporaneous lags
+# - Enhanced integration with causality_pred() and causality_predVAR() functions
+# - Support for restricted coefficient matrices in VAR models
+#
+# @keywords internal
+# @noRd
 
 VAR <- function(y, p = 1, type = c("const", "trend", "both", "none"),
     season = NULL, exogen = NULL, lag.max = NULL,

R/VARselect.R

@@ -1,10 +1,20 @@
-# This is a copy of the code from the package vars by Bernhard Pfaff
-# with changes that were not merged to the package
-# with the pull request from 2021-08-22
+# Internal VAR Order Selection Function (Modified from vars package)
+#
+# This is a modified copy of the VARselect() function from the vars package by Bernhard Pfaff.
+# The modifications were submitted via pull request on 2021-08-22 but were not merged:
 # https://github.com/bpfaff/vars/pull/10/files
-# I.e., subsequent vars v.1.5-6 (2021-09-17) does not include these changes.
-# The changes (most of them are marked with "#VL") allow for restricted VAR
-# estimation and Granger causality testing using such models.
+#
+# The vars package v.1.5-6 (2021-09-17) and later versions do not include these changes.
+# This modified version is maintained internally to support restricted VAR estimation
+# with lag restrictions for Granger causality testing.
+#
+# Key Modifications (marked with "#VL" comments):
+# - Added lag.restrict parameter for excluding near-contemporaneous lags
+# - Validation checks for lag.restrict parameter
+# - Integration with modified VAR() function
+#
+# @keywords internal
+# @noRd
 
 VARselect <- function(y,
                       lag.max = 10,

R/fun_causality.R

@@ -1,3 +1,16 @@
+# Internal Helper Functions for Granger Causality Testing
+#
+# This file contains internal (non-exported) helper functions for Granger causality
+# testing with restricted VAR models. These functions are used by causality_pred()
+# and causality_predVAR().
+#
+# Functions:
+#   - restrictions(): Creates/overlays restriction matrix for VAR models
+#   - caustests(): Computes multiple causality test statistics
+#
+# @keywords internal
+# @noRd
+
 # Function to recreate matrix of restrictions and overlay new restrictions for causality testing
 restrictions <- function(x, cause) {
     co.names <- vars::Bcoef(x)

R/fun_cusum.R

@@ -1,5 +1,16 @@
-#Auxiliary functions related to the modified CUSUM statistic 
-#by Horvath et al. (2017) for testing changepoints.
+# Internal Helper Functions for Modified CUSUM Changepoint Tests
+#
+# This file contains internal (non-exported) helper functions for implementing
+# the modified CUSUM statistic by Horvath et al. (2017) for changepoint detection.
+# These functions are used by mcusum_test() and related changepoint detection methods.
+#
+# Functions:
+#   - Mfun(): Computes M(k1,...,km) statistic from p. 553
+#   - M1fun(): Optimized version for at most one changepoint
+#   - MTfun(): Computes MT statistic from p. 554 or first equation on p. 568
+#
+# @keywords internal
+# @noRd
 
 
 ##### The function for M(k1,...,km) on p. 553:

R/fun_trust.R

@@ -1,5 +1,18 @@
-#Auxiliary functions related to the TRUST algorithm;
-#were in the main body of the package until v.5.
+# Internal Helper Functions for TRUST Clustering Algorithm
+#
+# This file contains internal (non-exported) helper functions for the TRUST
+# (TRUSTworthy Spatio-Temporal cluster detection) algorithm. These functions
+# were previously exported in package versions prior to v.5.
+# They are used by BICC(), CSlideCluster(), CWindowCluster(), and related clustering methods.
+#
+# Functions:
+#   - CNeighbor(): Identifies E^theta_delta neighbors of time series
+#   - CHomogeneity(): Checks homogeneity of time series clusters
+#   - CExpandSlideCluster(): Expands slide-level time series clusters
+#   - CExpandWindowCluster(): Expands window-level time series clusters
+#
+# @keywords internal
+# @noRd
 
 
 # ##### Neighborhood of Time Series #####

R/toMlm.R

@@ -1,10 +1,20 @@
-# This is a copy of the code from the package vars by Bernhard Pfaff
-# with changes that were not merged to the package
-# with the pull request from 2021-08-22
+# Internal VAR to MLM Conversion Function (Modified from vars package)
+#
+# This is a modified copy of toMlm() methods from the vars package by Bernhard Pfaff.
+# The modifications were submitted via pull request on 2021-08-22 but were not merged:
 # https://github.com/bpfaff/vars/pull/10/files
-# I.e., subsequent vars v.1.5-6 (2021-09-17) does not include these changes.
-# The changes (most of them are marked with "#VL") allow for restricted VAR
-# estimation and Granger causality testing using such models.
+#
+# The vars package v.1.5-6 (2021-09-17) and later versions do not include these changes.
+# This modified version is maintained internally to support conversion of restricted
+# VAR models to multivariate linear models (MLM) for coefficient estimation and testing.
+#
+# Key Modifications (marked with "#VL" comments):
+# - Support for x$restrictions matrix in varest objects
+# - Proper handling of restricted coefficients in conversion to MLM
+# - Individual equation fitting respecting coefficient restrictions
+#
+# @keywords internal
+# @noRd
 
 toMlm <- function(x, ...){
     UseMethod("toMlm")

README.md

@@ -23,6 +23,7 @@ library(funtimes)
 ## Current GitHub version
 
 - Added new function `r_crit()` to compute critical values of correlations.
+- In the function `causality_pred()`, added option to turn off automatic lag selection. 
 
 ## CRAN version 9.1
 

misc/README.md

@@ -0,0 +1,28 @@
+# Misc Directory
+
+This directory contains experimental, deprecated, or archived code that is not part of the main package.
+
+## Contents
+
+### Archived/Removed Functions
+
+- **TopoCBN.R** - Topological clustering function using Betti numbers. This function was removed from the package in version 8.2 due to removal of the `TDA` package dependency. The implementation is preserved here for reference.
+
+### Experimental/Incomplete Code
+
+- **CRAD.R** and **CRAD.py** - Experimental clustering using robust anomaly detection. This appears to be incomplete/under development and is not currently integrated into the package.
+- **CRAD.Rd** - Draft documentation for CRAD function.
+
+### Old Versions
+
+- **causality_predVAR_v0.R** - Previous version of the `causality_predVAR()` function, retained for reference during development.
+
+## Recommendations
+
+1. **CRAD files** - Consider either:
+   - Completing the implementation and integrating into the package, or
+   - Removing if the approach was abandoned
+
+2. **TopoCBN.R** - Keep for archival purposes since it was a documented feature removed for dependency reasons. Users may still want to access it.
+
+3. **causality_predVAR_v0.R** - Can be removed once the new version is stable and well-tested.