#include <liarsdice/statistics/time_series.hpp>
template<typename T, std::size_t MaxSize = 1000>
liarsdice::statistics::TimeSeries class

Time series container using boost::circular_buffer.

Template parameters
T Value type
MaxSize Maximum number of points to store

Efficient storage for time-ordered data with automatic old data removal. Supports various time series analysis operations including moving averages, trend detection, and seasonal decomposition.

Public types

using value_type = T
using time_point = TimePoint<T>
using duration = std::chrono::steady_clock::duration
using buffer_type = boost::circular_buffer<time_point>

Constructors, destructors, conversion operators

TimeSeries()
Default constructor.

Public functions

void add(const T& value)
Add a value with current timestamp.
void add(std::chrono::steady_clock::time_point timestamp, const T& value)
Add a value with specific timestamp.
auto size() const -> std::size_t
Get size of time series.
auto empty() const -> bool
Check if empty.
auto latest() const -> boost::optional<T>
Get most recent value.
auto oldest() const -> boost::optional<T>
Get oldest value.
auto get_window(duration window_duration) const -> std::vector<T>
Get values within time window.
auto simple_moving_average(std::size_t periods) const -> std::vector<double>
Calculate simple moving average.
auto exponential_moving_average(double alpha) const -> std::vector<double>
Calculate exponential moving average.
auto linear_trend() const -> std::pair<double, double>
Detect trend using linear regression.
auto rate_of_change(std::size_t periods) const -> boost::optional<double>
Calculate rate of change.
auto detect_outliers(double threshold = 3.0) const -> std::vector<std::size_t>
Detect outliers using z-score method.
auto savitzky_golay_filter(std::size_t window_size, std::size_t poly_order) const -> std::vector<double>
Apply Savitzky-Golay filter for smoothing.
auto autocorrelation(std::size_t lag) const -> boost::optional<double>
Calculate autocorrelation.
auto get_data() const -> std::vector<time_point>
Get raw data points.
void clear()
Clear all data.
auto resample(duration interval) const -> TimeSeries
Resample time series to fixed intervals.

Function documentation

template<typename T, std::size_t MaxSize>
void liarsdice::statistics::TimeSeries<T, MaxSize>::add(const T& value)

Add a value with current timestamp.

Parameters
value Value to add

template<typename T, std::size_t MaxSize>
void liarsdice::statistics::TimeSeries<T, MaxSize>::add(std::chrono::steady_clock::time_point timestamp, const T& value)

Add a value with specific timestamp.

Parameters
timestamp Time point
value Value to add

template<typename T, std::size_t MaxSize>
std::size_t liarsdice::statistics::TimeSeries<T, MaxSize>::size() const

Get size of time series.

Returns Number of data points

template<typename T, std::size_t MaxSize>
bool liarsdice::statistics::TimeSeries<T, MaxSize>::empty() const

Check if empty.

Returns True if no data points

template<typename T, std::size_t MaxSize>
boost::optional<T> liarsdice::statistics::TimeSeries<T, MaxSize>::latest() const

Get most recent value.

Returns Optional containing last value

template<typename T, std::size_t MaxSize>
boost::optional<T> liarsdice::statistics::TimeSeries<T, MaxSize>::oldest() const

Get oldest value.

Returns Optional containing first value

template<typename T, std::size_t MaxSize>
std::vector<T> liarsdice::statistics::TimeSeries<T, MaxSize>::get_window(duration window_duration) const

Get values within time window.

Parameters
window_duration Duration of window from now
Returns Vector of values within window

template<typename T, std::size_t MaxSize>
std::vector<double> liarsdice::statistics::TimeSeries<T, MaxSize>::simple_moving_average(std::size_t periods) const

Calculate simple moving average.

Parameters
periods Number of periods
Returns Vector of moving averages

template<typename T, std::size_t MaxSize>
std::vector<double> liarsdice::statistics::TimeSeries<T, MaxSize>::exponential_moving_average(double alpha) const

Calculate exponential moving average.

Parameters
alpha Smoothing factor (0 < alpha < 1)
Returns Vector of EMAs

template<typename T, std::size_t MaxSize>
std::pair<double, double> liarsdice::statistics::TimeSeries<T, MaxSize>::linear_trend() const

Detect trend using linear regression.

Returns Pair of (slope, intercept)

template<typename T, std::size_t MaxSize>
boost::optional<double> liarsdice::statistics::TimeSeries<T, MaxSize>::rate_of_change(std::size_t periods) const

Calculate rate of change.

Parameters
periods Number of periods to look back
Returns Rate of change percentage

template<typename T, std::size_t MaxSize>
std::vector<std::size_t> liarsdice::statistics::TimeSeries<T, MaxSize>::detect_outliers(double threshold = 3.0) const

Detect outliers using z-score method.

Parameters
threshold Z-score threshold (default: 3)
Returns Indices of outliers

template<typename T, std::size_t MaxSize>
std::vector<double> liarsdice::statistics::TimeSeries<T, MaxSize>::savitzky_golay_filter(std::size_t window_size, std::size_t poly_order) const

Apply Savitzky-Golay filter for smoothing.

Parameters
window_size Window size (must be odd)
poly_order Polynomial order
Returns Smoothed values

template<typename T, std::size_t MaxSize>
boost::optional<double> liarsdice::statistics::TimeSeries<T, MaxSize>::autocorrelation(std::size_t lag) const

Calculate autocorrelation.

Parameters
lag Lag value
Returns Autocorrelation coefficient

template<typename T, std::size_t MaxSize>
std::vector<time_point> liarsdice::statistics::TimeSeries<T, MaxSize>::get_data() const

Get raw data points.

Returns Vector of time points

template<typename T, std::size_t MaxSize>
TimeSeries liarsdice::statistics::TimeSeries<T, MaxSize>::resample(duration interval) const

Resample time series to fixed intervals.

Parameters
interval Target interval duration
Returns Resampled time series