# Import used libraries library(PeakSegOptimal) library(PeakSegDP) library(microbenchmark) library(ggplot2)
# Initialize n_seq (different values of n to test with) n_seq = c(10, 100, 1000, 10000) # Declare a vector to store the runtime data in each loop iteration for each algorithm cdpa_data <- integer(length(n_seq)) pdpa_data <- integer(length(n_seq))
# Loop through the different values of input N and compute the benchmark in each iteration
for (i in seq(1, length(n_seq))){
# GeneratePoisson distribution
x <- rpois(n_seq[i], 10)
# Benchmark both the PeakSegPDPA and cDPA functions with maxSegments = 3
m <- summary(microbenchmark( PeakSegPDPA(x, rep(1, length(x)), 3L),cDPA(x, rep(1, length(x)), 3L)))
pdpa_data[i] <- m$mean[1]
cdpa_data[i] <- m$mean[2]
}
# Create a dateframe from the computed data df = data.frame(pdpa_data, cdpa_data, n_seq) # Plot cDPA vs N(red) and PeakSegPDPA vs N(blue) ggplot(df, aes(x=n_seq, y=cdpa_data)) + geom_line(color = 'red') + geom_line(y = pdpa_data, color='blue') + labs(x="N", y="Runtime")
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4
