Missing dates

RyanRezel

Hi,
I have a been given a dataset to predict employee termination rate, I have one attribute that is the date of termination and another that is Binominal and states if the employee has been terminated in the past or not if he/she has not then in the date of termination it is identified as a missing value. further, in the date of termination attribute the date is in two forms, for example, 11/02/15 and 12/05/2012 and therefore RM identifies it as polynomial and hence I cannot convert it into a date and I don't know how to treat these missing values.
I wanted to insert something like N/A in the missing values but that is I reckon not advisable.
In such a case is it advisable to use one of these as my label?
what is my best course of action, I hope I can get a quick response, thank you community.


cheers
Ryan

MartinLiebig

Hi,

depends on how malformed they are. One thing would be to filter them out up front. Or to replace them with a missing using Generate attributes with

if(matches(...),MISSING_NOMINAL, date)

or so.

Best,

Martin

MartinLiebig

Hi,

did you try Polynominal to Date?

Best,

Martin

RyanRezel

Yeah that’s Ive tried That and it gives me an error message saying “cannot parse date”

cheers
Ryan

MartinLiebig

Hi,

the attached process works for me? Maybe you have malformed dates in your data set?

BR,

Martin

<宏/ >

<参数键= "柯ep_old_attribute" value="false"/>






> < /过程

> < /过程

RyanRezel

Yes I do!
how do I deal with malformed dates?