Error in iv.str(df, "tmp_iv_calc_label", y) : Not a binary outcome about woe HOT 9 OPEN

hello~I can't install the package successfully by running the following code,could you please give me some help?

library(devtools)
install_github("tomasgreif/woe")
library(woe)
the error is like this:
install_github("tomasgreif/woe")
Downloading github repo tomasgreif/woe@master
Error in curl::curl_fetch_memory(url, handle = handle) :
Timeout was reached

from woe.

I have the same issue @kevinogoro, all of my numerical variables would yield this error.

from woe.

Same problem as @kevinogoro and @Chuck-UGA. The german_data works fine, but my data crashes. What's wrong?

from woe.

Hi @XiaoxiaoWang87 @Chuck-UGA . I am not mantaining this package as it was more an exercise in package development. If you can send me a dataset where error can be reproduced I can try to find the source of error.

from woe.

Hi @tomasgreif , all. Thanks! I eventually solved my problem by doing row.names(dat) <- 1:nrow(dat) before running iv.mult(). I'm not sure why but originally my dataset row names are not ordered, which potentially caused the crash. Perhaps other people had a similar problem?

from woe.

Yeah, that could cause a problem. When nodes from decision trees are
assigned back to data the algorithm silently assumes that rownames are
ordered.

On 10 August 2015 at 20:34, XiaoxiaoWang87 [email protected] wrote:

Hi @tomasgreif https://github.com/tomasgreif , all. Thanks! I
eventually solved my problem by doing row.names(dat) <- 1:nrow(dat) before
running iv.mult(). I'm not sure why but originally my dataset row names are
not ordered, which potentially caused the crash. Perhaps other people had a
similar problem?

—
Reply to this email directly or view it on GitHub
#1 (comment).

from woe.

I had a same issue. After debugging, I found out the issue caused by unmatched row names. Since I deleted some rows before, but in R the row.names won't update. For example, if I delete 10th row, the row.names would show 7,8,9,11. In this package, in function ivr.num(), it uses row.names to be 'foreign key' to merge two data frame, so causes NA rows. At the end, outcome variable Y would have 3 classes, which are '0', '1', '' and shows the error message: iv.str(df, "tmp_iv_calc_label", y) : Not a binary outcome. In order to solve the problem, you can reorder row names of your data before use the function by doing row.names(data) = seq(1,nrow(data)). Hope it helps!

from woe.

One could just do the rownames(dataset)=NULL, the algo will start working.

from woe.

can this function -> iv.mult(train_data, "ABOVE50K", TRUE), be used when my train_data has both character and integer data

from woe.