Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

about NA values and featurebased complexity #64

Open
isilay123 opened this issue Mar 28, 2023 · 6 comments
Open

about NA values and featurebased complexity #64

isilay123 opened this issue Mar 28, 2023 · 6 comments

Comments

@isilay123
Copy link

Hi,
Thank you very much for this useful work. I have two questions.

  1. Is it not possible to calculate feature-based or other complexity measures for each data? Sometimes, I see this error:Error in match.arg(measures, ls.correlation(), TRUE) :
    'arg' should be one of “C1”, “C2”, “C3”, “C4”

  2. I couldn't understand the NA output reason. For example, when I recognize the NA value in N3, can I say it doesn't have a standard deviation? It calculates both mean and standard deviation, but do these metrics have standard deviation values because they calculate more than one mean?

Kind regards,
Isill

@LesleyWheat
Copy link

Not the developer. For point 2, N3 should not return NA for sd as that would indicate the leave one out error is only one value (sd of a one item vector is NA in R). Have you checked that your input values are formatted correctly? Also, there is always only one mean, just multiple values for leave one out error. Ex. an error rate of 56 out of 495 points is mean 0.113 and sd of 0.317.

@isilay123
Copy link
Author

Not the developer. For point 2, N3 should not return NA for sd as that would indicate the leave one out error is only one value (sd of a one item vector is NA in R). Have you checked that your input values are formatted correctly? Also, there is always only one mean, just multiple values for leave one out error. Ex. an error rate of 56 out of 495 points is mean 0.113 and sd of 0.317.

Thank you for your answer. I will check again, but I couldn't understand your example, for example, I calculated
featurebased.F1.mean = 2.775642e-01 , featurebased.F1.sd 2.612623e-01 but I couldn't understand.

@LesleyWheat
Copy link

Not the developer. For point 2, N3 should not return NA for sd as that would indicate the leave one out error is only one value (sd of a one item vector is NA in R). Have you checked that your input values are formatted correctly? Also, there is always only one mean, just multiple values for leave one out error. Ex. an error rate of 56 out of 495 points is mean 0.113 and sd of 0.317.

Thank you for your answer. I will check again, but I couldn't understand your example, for example, I calculated featurebased.F1.mean = 2.775642e-01 , featurebased.F1.sd 2.612623e-01 but I couldn't understand.

I'm not sure exactly what the sd value for F1 is, I work primarily with the neighbourhood measures. I was talking about N3, the leave one out error rate can be represented as a vector of classified correct/incorrectly as 0 and 1, from that the mean and sd can be taken. I don't personally use that sd, but it seems strange to be NA.

@isilay123
Copy link
Author

Thank you for answer

@aclorena
Copy link
Collaborator

aclorena commented Mar 30, 2023 via email

@isilay123
Copy link
Author

The original F1 measure takes only the maximum discriminant value among all features. For obtaining this result, you should choose max as a summarization function of the measure. The mean and std are output when considering the discriminant value of all features. Sent with Right Inbox https://www.rightinbox.com/?utm_source=signature Em qua., 29 de mar. de 2023 às 17:54, isilay123 @.> escreveu:

Thank you for answer — Reply to this email directly, view it on GitHub <#64 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFIURWTOHGQD2X42GVTTXHLW6SORPANCNFSM6AAAAAAWKSTJYE . You are receiving this because you are subscribed to this thread.Message ID: @.
>
-- Prof Ana Carolina Lorena Divisão de Ciência da Computação (IEC) Instituto Tecnológico de Aeronáutica (ITA)

Thank you very much .

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants