| Title: | Quantum Time Series Analysis: Drift, Noise Spectroscopy and Calibration Forecasting |
| Version: | 0.1.1 |
| Author: | Leila Marvian Mashhad [aut, cre] |
| Maintainer: | Leila Marvian Mashhad <leila.marveian@gmail.com> |
| Description: | Tools for exploratory statistical analysis of quantum-hardware calibration time series. The package provides simulators for random telegraph noise (RTN), power-law noise, and Ornstein-Uhlenbeck dephasing; Welch and sine-multitaper power spectral density estimators; a lightweight two-state hidden Markov model for switching signals; cumulative sum (CUSUM) and binary-segmentation diagnostics for calibration drift; residual-quantile interval forecasts; and filter-function calculations for illustrative coherence curves. The package includes a reproducible generator of simulated superconducting-qubit calibration records; it does not retrieve authenticated live provider data. Methodological background is provided by Welch (1967) <doi:10.1109/TAU.1967.1161901>, Thomson (1982) <doi:10.1109/PROC.1982.12433>, Rabiner (1989) <doi:10.1109/5.18626>, Page (1954) <doi:10.1093/biomet/41.1-2.100>, Paladino et al. (2014) <doi:10.1103/RevModPhys.86.361>, and Cywinski et al. (2008) <doi:10.1103/PhysRevB.77.174509>. |
| License: | MIT + file LICENSE |
| Encoding: | UTF-8 |
| RoxygenNote: | 7.3.2 |
| Depends: | R (≥ 4.1.0) |
| Imports: | stats, graphics, utils |
| NeedsCompilation: | no |
| Packaged: | 2026-07-28 17:03:42 UTC; user |
| Repository: | CRAN |
| Date/Publication: | 2026-08-06 13:10:09 UTC |
Compute an illustrative coherence value from a PSD
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
compute_Wt(psd, filter)Arguments
psd |
data frame with f and S columns |
filter |
output from filter_function_q |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A list with chi, W, interpolated spectrum, filter values, and frequency.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Binary-segmentation drift diagnostic
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
cp_qdrift(x, pen = NULL, min_seg = 20)Arguments
x |
numeric time series |
pen |
penalty for adding a changepoint |
min_seg |
minimum segment length |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A list with changepoints, segment means, and the number of segments.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
CUSUM drift diagnostic
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
cusum_q(x, target = NULL, k = NULL, h = NULL, sigma = NULL)Arguments
x |
numeric time series |
target |
in-control target mean; default uses early observations |
k |
reference value |
h |
decision interval |
sigma |
in-control standard deviation |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A list containing CUSUM paths, parameter values, and alarm indices.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Compute an illustrative decoherence curve
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
decoherence_curve(psd, t_grid = seq(0.1, 10, length.out = 100), n_pulses = 1)Arguments
psd |
data frame with f and S columns |
t_grid |
times at which to evaluate coherence |
n_pulses |
number of idealized pi pulses |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A data frame with t, W, and chi.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Estimate a two-state switching model
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
estimate_rtn_hmm(x, n_states = 2, max_iter = 50, seed = NULL)Arguments
x |
numeric time series |
n_states |
number of states; the implementation is intended for two states |
max_iter |
maximum number of hard-EM refinement iterations |
seed |
optional random-number seed for k-means initialization; |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A list containing state means, standard deviations, transition matrix, decoded states, dwell summaries, and amplitude.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Generate simulated IBM-like calibration data
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
fetch_ibm_calib(n = 500, backend = "ibmq_manila_sim", seed = NULL, simulate = TRUE)Arguments
n |
number of time points |
backend |
label included in the output; no network request is made |
seed |
optional random-number seed; |
simulate |
retained for backward compatibility; only simulated data are produced |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A data frame with time, backend, T1, T2, freq_GHz, and readout_err columns.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Compute a simple fidelity proxy
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
fidelity_ts(T1, T2, gate_time = 0.05)Arguments
T1 |
numeric relaxation-time series |
T2 |
numeric dephasing-time series |
gate_time |
gate duration in the same time units as T1 and T2 |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A numeric vector of bounded fidelity-proxy values.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Construct a simple dynamical-decoupling filter function
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
filter_function_q(t, n_pulses = 1, f, dt = NULL)Arguments
t |
total evolution time |
n_pulses |
number of idealized pi pulses |
f |
numeric frequency vector |
dt |
reserved time-resolution argument |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A list with frequency, angular frequency, filter values, time, and pulse count.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Alias for forecast_calib
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
forecast_T1(x, h = 20, window = 100, alpha = 0.1, method = "ets")Arguments
x |
numeric T1 series |
h |
forecast horizon |
window |
rolling window length |
alpha |
nominal miscoverage level |
method |
forecast method |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A list returned by forecast_calib.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Rolling residual interval forecast
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
forecast_calib(x, h = 20, window = 100, alpha = 0.1, method = "ets")Arguments
x |
numeric time series |
h |
forecast horizon |
window |
length of the rolling training window |
alpha |
nominal miscoverage level used for the residual quantile |
method |
one of "mean", "ets", or "arima"; the current implementation uses lightweight baseline forecasts |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A list with point forecast, lower and upper interval limits, residuals, and residual quantile.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Alias for estimate_rtn_hmm
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
hmm_rtn_fit(x, n_states = 2, max_iter = 50, seed = NULL)Arguments
x |
numeric time series |
n_states |
number of states |
max_iter |
maximum number of iterations |
seed |
optional random-number seed for k-means initialization; |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A list returned by estimate_rtn_hmm.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Plot calibration drift diagnostics
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
plot_drift(x, metric = "T1", ...)Arguments
x |
data frame returned by fetch_ibm_calib or a numeric series |
metric |
column to plot when x is a data frame |
... |
additional arguments reserved for future use |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
Invisibly, a list containing CUSUM and changepoint diagnostics.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Alias for psd_multitaper_q
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
psd_mtm_q(x, fs = 1, nw = 3, k = 5)Arguments
x |
numeric time series |
fs |
sampling frequency in Hz |
nw |
nominal time-bandwidth parameter |
k |
number of sine tapers |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A data frame with f, S, S_low, and S_high.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Sine-multitaper power spectral density estimator
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
psd_multitaper_q(x, fs = 1, nw = 3, k = 5)Arguments
x |
numeric time series |
fs |
sampling frequency in Hz |
nw |
nominal time-bandwidth parameter used to choose a conventional taper count |
k |
number of sine tapers |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A data frame with f, S, S_low, and S_high.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Welch power spectral density estimator
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
psd_welch(x, fs = 1, nperseg = 256, noverlap = NULL, window = "hann")Arguments
x |
numeric time series |
fs |
sampling frequency in Hz |
nperseg |
segment length |
noverlap |
number of overlapping observations |
window |
window type; currently "hann" is supported |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A data frame with frequency f and one-sided spectral density S.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Power-law noise simulator
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
r_oneoverf(n, alpha = 1, fs = 1, f_low = NULL)Arguments
n |
positive number of samples |
alpha |
power-law exponent |
fs |
sampling frequency in Hz |
f_low |
optional low-frequency cutoff |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A list with time vector, simulated signal, sampling frequency, and exponent.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Ornstein-Uhlenbeck dephasing simulator
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
r_ou_dephasing(n, tau_c = 10, sigma = 1, dt = 1)Arguments
n |
positive number of samples |
tau_c |
correlation time |
sigma |
stationary standard deviation |
dt |
time increment |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A numeric vector containing a simulated process.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.
Random Telegraph Noise simulator
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
r_rtn(n, fs = 1, amplitude = 1, tau_up = 50, tau_down = 50, p0 = 0.5)Arguments
n |
positive number of samples |
fs |
sampling frequency in Hz |
amplitude |
difference between the two signal levels |
tau_up |
mean dwell time in the upper state, in samples |
tau_down |
mean dwell time in the lower state, in samples |
p0 |
probability of starting in the upper state |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
A numeric vector containing the simulated RTN signal.
Author(s)
Leila Marvian Mashhad
Examples
set.seed(1)
x <- r_rtn(100)
head(x)
Viterbi decoder for a switching model
Description
Part of the qtsa toolkit for exploratory analysis of quantum-hardware calibration time series.
Usage
viterbi_rtn(log_emission, log_trans)Arguments
log_emission |
matrix of log emission probabilities with observations in rows |
log_trans |
matrix of log transition probabilities |
Details
This function is intended for exploratory analysis. Interpret physically meaningful results only after checking sampling frequency, units, and model assumptions.
Value
An integer vector giving the most likely state path.
Author(s)
Leila Marvian Mashhad
Examples
x <- rnorm(100)
# See the package vignette or help pages for a complete workflow.