I created an algorithm and I tested it against a current algorithm.

The results are in this form:

Power   Processes   Method  Time(s)
1          3          1     19,94
1          4          1     20,04
1          5          1     20,06
1          6          1     19,95
1          7          1     20,1
1          8          1     20,03
1          3          0     30,3 

for each method where my method is "1" and the other message is represented by "0". Process indicates the available processing power (I only have 4 servers therefore 3,4 processes may run each on a single server, 5-8 servers have to share resources - not indicated in the example table)

I've made 10 replications each test.

I wanted to create a linear regression comparing both models in order to show that even with more processes my algorithm runs faster. But the graphs and statics I could generate with ANOVA didn't really help me?

Which methods do I have to use? And how may I generate graphs explaining the differences between the regressions?


Creating the Model:

LinearModel.1 <- lm(Time.s. ~ Processes*Method+Power, data=Dataset)

show different slopes:

scatterplot(Time.s.~Processes | MethodF, log="y", reg.line=lm,
smooth=FALSE,    spread=FALSE, boxplots=FALSE, span=0.5,
ellipse=FALSE, levels=c(.5, .9),    by.groups=TRUE, data=Dataset)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.