The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
dd_decompose() – flow decomposition of each change in
the decision rate into inflow (uptake) and outflow (reversal) using the
marginal-transition identity. Reports turnover, net endpoint change
(NEC), cumulative marginal burden (CMB), cumulative transition burden
(CTB), and, for unbalanced panels, the gap between paired and
cross-sectional change.dd_group_decompose() – symmetric midpoint decomposition
of aggregate change into within-group and composition components,
composition drift D_C, and group-specific uptake and
reversal. Supports time-varying group membership.dd_cluster_bootstrap() – unit-level cluster bootstrap
that keeps each unit’s full history. One set of draws can be shared by
all inference functions.dd_global_test() – global test that uptake, reversal,
and/or prevalence are constant across waves, overall or by group. Max-t
and Wald statistics with a centred bootstrap null; familywise-adjusted
p-values and simultaneous confidence bands. Contrasts compare each step
with the previous one (contrast = "consecutive", default)
or with the first one (contrast = "baseline", more power
against gradual drift).dd_simulate() – simulate panels under stable, marginal,
masked, abrupt, group-masked, and composition-change scenarios, or
custom uptake and reversal sequences.inst/simulation/mc_study.R – script for the Monte Carlo
study comparing the prevalence trend test with the global stability
test.dd_decompose, dd_group_decompose, and
dd_global_test have print() and
plot() methods.dd_audit() no longer returns
"no drift detected", "marginal drift",
"moderate drift", or "strong drift". Those
labels came from fixed, uncalibrated cut-offs. The audit now runs one
cluster bootstrap, the flow and group decompositions, and
dd_global_test(), and returns a conclusion
stating the test result at level alpha.
verdict is kept as an alias of conclusion. New
arguments: global_test, components,
alpha, thresholds (user-supplied practical
cut-offs), and seed.dd_changepoint(): the event-alignment result no longer
has a "strong"/"weak" evidence
label; it returns alpha and a logical
significant. Documentation now notes that the wave-level
methods have little power with few waves.dd_transition(): reversal_drift held the
mean absolute change in uptake P(1|0), not reversal. It now holds the
mean absolute change in reversal P(0|1). The uptake value is now
returned as uptake_drift. TDI values are unchanged. Code
that used reversal_drift from 0.1.0 was reading uptake
drift and should switch to uptake_drift to keep the old
numbers.dd_transition() and dd_indices()
documentation now gives the TDI and CDB formulas in terms of uptake,
reversal, and the decision rate.plot.dd_transition() now shows uptake and
reversal.dd_build() — construct a drift_panel
object from long-format panel data.dd_prevalence() — detect drift in decision rate;
computes Decision Drift Index (DDI).dd_transition() — detect drift in transition structure;
computes Transition Drift Index (TDI).dd_entropy_trend() — track rolling Shannon entropy and
switching rate over time.dd_group_drift() — detect group-differential drift;
computes Group Differential Drift (GDD).dd_changepoint() — change-point and regime-shift
detection via CUSUM, segmented regression, and event alignment.dd_indices() — compute all four original drift indices
in one call: DDI, TDI, GDD (Group Differential Drift), CDB (Cumulative
Drift Burden).dd_robustness() — test stability of drift conclusions
across analytic choices (balanced panel, leave-one-period-out,
leave-one-group-out, min_waves grid, bootstrap CI).dd_sensitivity() — probe vulnerability to data
problems: decision miscoding, missing-wave attrition, threshold shifts,
and subgroup composition shifts.dd_audit() — integrated three-layer audit (detection →
decomposition → stress-testing) with a single-call verdict.plot() methods for all module outputs.plot.dd_audit() (requires
patchwork).decisionpathsdecisionpaths asks: What happened? (path
construction, DRI, entropy, equity)DecisionDrift asks: Did the system change?
(drift detection, decomposition, stress-testing)These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.