как решить ошибки млр, роз и карет в г - PullRequest
Новая
       12

как решить ошибки млр, роз и карет в г

0 голосов
/ 03 июля 2018

Я работаю над дисбалансом, поэтому я пытаюсь сбалансировать dataset, и для этого я попробовал различные методы, такие как caret, mlr, ROSE, но получил ошибку?

ул (MYDATA) 

Classes ‘data.table’ and 'data.frame':  40596053 obs. of  8 variables:
 $ SessionID  : chr  "1" "1" "1" "1" ...
 $ Timestamp.x: chr  "2014-04-07T10:51:09.277Z" "2014-04-07T10:54:09.868Z" "2014-04-07T10:54:46.998Z" "2014-04-07T10:57:00.306Z" ...
 $ ItemID.x   : chr  "214536502" "214536500" "214536506" "214577561" ...
 $ Category   : chr  "0" "0" "0" "0" ...
 $ Timestamp.y: chr  NA NA NA NA ...
 $ ItemID.y   : chr  "0" "0" "0" "0" ...
 $ Price      : chr  NA NA NA NA ...
 $ Quantity   : chr  NA NA NA NA ...
 - attr(*, ".internal.selfref")=<externalptr> 
 - attr(*, "sorted")= chr "SessionID"

dput (head (mydata, 20)) 

structure(list(SessionID = c("1", "1", "1", "1", "10000001", 
"10000001", "10000001", "10000001", "10000001", "10000002", "10000002", 
"10000002", "10000002", "10000003", "10000003", "10000003", "10000004", 
"10000004", "10000004", "10000004"), Timestamp.x = c("2014-04-07T10:51:09.277Z", 
"2014-04-07T10:54:09.868Z", "2014-04-07T10:54:46.998Z", "2014-04-07T10:57:00.306Z", 
"2014-09-08T10:35:38.841Z", "2014-09-08T10:40:20.143Z", "2014-09-08T10:40:36.704Z", 
"2014-09-08T10:41:12.386Z", "2014-09-08T10:48:34.245Z", "2014-09-08T19:10:51.206Z", 
"2014-09-08T19:13:31.104Z", "2014-09-08T19:14:54.518Z", "2014-09-08T19:33:38.355Z", 
"2014-09-05T11:32:15.524Z", "2014-09-05T11:34:25.159Z", "2014-09-05T11:37:23.321Z", 
"2014-09-05T13:14:45.867Z", "2014-09-05T13:14:45.867Z", "2014-09-05T13:55:18.886Z", 
"2014-09-05T13:55:18.886Z"), ItemID.x = c("214536502", "214536500", 
"214536506", "214577561", "214854230", "214556216", "214556212", 
"214854230", "214854125", "214849322", "214838094", "214714721", 
"214853711", "214853090", "214851326", "214853094", "214853090", 
"214853090", "214851326", "214851326"), Category = c("0", "0", 
"0", "0", "S", "S", "S", "S", "S", "S", "S", "S", "S", "3", "3", 
"3", "3", "3", "3", "3"), Timestamp.y = c(NA, NA, NA, NA, NA, 
NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, "2014-09-05T14:07:33.845Z", 
"2014-09-05T14:07:33.845Z", "2014-09-05T14:07:33.845Z", "2014-09-05T14:07:33.845Z"
), ItemID.y = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L), .Label = c("0", 
"1"), class = "factor"), Price = c(NA, NA, NA, NA, NA, NA, NA, 
NA, NA, NA, NA, NA, NA, NA, NA, NA, "4188", "1046", "4188", "1046"
), Quantity = c(NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, 
NA, NA, NA, NA, "1", "1", "1", "1")), .Names = c("SessionID", 
"Timestamp.x", "ItemID.x", "Category", "Timestamp.y", "ItemID.y", 
"Price", "Quantity"), sorted = "SessionID", class = c("data.table", 
"data.frame"), row.names = c(NA, -20L), .internal.selfref = <pointer: 0x0000000013950788>)

с использованием млр 

task = makeClassifTask(data = mydata, target = "ItemID.y") 
Warning in makeTask(type = type, data = data, weights = weights, blocking = blocking,  :
  Provided data is not a pure data.frame but from class data.table, hence it will be converted.
Error in (function (cn, x)  : 
  Unsupported feature type (character) in column 'SessionID'.

поэтому я изменил тип sessionID с char на num, но получил ту же ошибку с Timestamp.x

Warning in makeTask(type = type, data = data, weights = weights, blocking = blocking,  :
  Provided data is not a pure data.frame but from class data.table, hence it will be converted.
Error in (function (cn, x)  : 
  Unsupported feature type (chr) in column 'Timestamp.x'.

Затем я изменил тип Timestamp.x с char на date, но снова получил ту же ошибку. 

Warning in makeTask(type = type, data = data, weights = weights, blocking = blocking,  :
      Provided data is not a pure data.frame but from class data.table, hence it will be converted.
    Error in (function (cn, x)  : 
      Unsupported feature type (Date) in column 'Timestamp.x'.

Я тоже пробовал карету и РОЗУ 

library(caret) 
> x <- matrix(mydata %>% select(-ItemID.y)
> y <- as.factor(mydata$ItemID.y)
> imbl_crt <- downSample(x, y, yname = "ItemID.y")
Error in `$<-.data.frame`(`*tmp*`, .outcome, value = c(1L, 1L, 1L, 1L,  : 
  replacement has 40596053 rows, data has 7

с использованием ROSE 

> library(ROSE)
> data_balanced_over <- ovun.sample(mydata$ItemID.y ~ ., data = mydata, method = "over", N = 40596053) #balance dataset using ROSE 
Error in (function (formula, data, method, subset, na.action, N, p = 0.5,  : 
  The response variable has only one class.

Как я исправил одну из этих проблем и сбалансировал dataset? Спасибо

...