roll_sum: Calculate the rolling sums over a _time series_ or a _matrix_...

In algoquant/HighFreq: High Frequency Time Series Management

Calculate the rolling sums over a time series or a matrix using Rcpp.

Description

Calculate the rolling sums over a time series or a matrix using Rcpp.

Usage

roll_sum(tseries, lookb = 1L, weightv = 0L)

Arguments

tseries	A time series or a matrix.
lookb	The length of the look-back interval, equal to the number of data points included in calculating the rolling sum (the default is lookb = 1).
weightv	A single-column matrix of weights (the default is weightv = 0).

Details

The function roll_sum() calculates the rolling weighted sums over the columns of the data tseries.

If the argument weightv is equal to zero (the default), then the function roll_sum() calculates the simple rolling sums of the time series data p_t over the look-back interval \Delta:

\bar{p}_t = \sum_{j=(t-\Delta+1)}^{t} p_j

If the weightv argument has the same number of rows as the argument tseries, then the function roll_sum() calculates rolling weighted sums of the time series data p_t in two steps.

It first calculates the rolling sums of the products of the weights w_t times the time series data p_t over the look-back interval \Delta:

\bar{w}_t = \sum_{j=(t-\Delta+1)}^{t} w_j

\bar{p}^w_t = \sum_{j=(t-\Delta+1)}^{t} w_j p_j

It then calculates the rolling weighted sums \bar{p}_t as the ratio of the sum products of the weights and the data, divided by the sums of the weights:

\bar{p}_t = \frac{\bar{p}^w_t}{\bar{w}_t}

The function roll_sum() returns a matrix with the same dimensions as the input argument tseries.

The function roll_sum() is written in C++ Armadillo code, so it's much faster than equivalent R code.

Value

A matrix with the same dimensions as the input argument tseries.

Examples

## Not run: 
# Calculate historical returns
retp <- na.omit(rutils::etfenv$returns[, c("VTI", "IEF")])
# Define parameters
lookb <- 11
# Calculate rolling sums and compare with rutils::roll_sum()
sumc <- HighFreq::roll_sum(retp, lookb)
sumr <- rutils::roll_sum(retp, lookb)
all.equal(sumc, coredata(sumr), check.attributes=FALSE)
# Calculate rolling sums using R code
sumr <- apply(zoo::coredata(retp), 2, cumsum)
sumlag <- rbind(matrix(numeric(2*lookb), nc=2), sumr[1:(NROW(sumr) - lookb), ])
sumr <- (sumr - sumlag)
all.equal(sumc, sumr, check.attributes=FALSE)
# Calculate weights equal to the trading volumes
weightv <- quantmod::Vo(rutils::etfenv$VTI)
weightv <- weightv[zoo::index(retp)]
# Calculate rolling weighted sums
sumc <- HighFreq::roll_sum(retp, lookb, 1/weightv)
# Plot dygraph of the weighted sums
datav <- cbind(retp$VTI, sumc[, 1])
colnames(datav) <- c("VTI", "Weighted")
endd <- rutils::calc_endpoints(datav, interval="weeks")
dygraphs::dygraph(cumsum(datav)[endd], main=colnames(datav)) %>% 
  dyOptions(colors=c("blue", "red"), strokeWidth=2) %>% 
  dyLegend(width=300)

## End(Not run)

algoquant/HighFreq documentation built on Oct. 26, 2024, 9:20 p.m.