Hello everyone, I need to ask how to retrieve the dataset to original string (nominal) after converting it to numerical, our data dataset about some diseases that component from row header and value in the instance so that have a class label, anyway not important the detail, I convert it to numerical to get better result, Particularly I'll implement the association rule algorithm, then the input data must be clear as string (nominal) to get the rule in final.
if the algorithm produces the rule that describes the relationship in numerical, I do not benefit from them, I want them to be in the string (nominal).
any suggestion to me about to retrieve the rule in numeric to nominal after generating the rule in numeric
Hello. What do you use as programmation language?
I work in this field. I use java and package weka that contains functions which facilitate coding and encoding. The idea, in general is explaind using this example.
If you have 4 values of a class, you can store them in a table in which indices are 0, 1, 2 and 3. For conversion, you use these indices. In inverse. You can obtain directly the corresponding nominal value.
Goog luck.
Aouatef Mahani i'm using weka tool and weka API in java . what the name of function that code and encoding ?
Hi Mohammed Thamer ,
First of all, in your question, you are considering string and nominal are from the same type. This is not accurate because string and nominal are different.
If your data from the numeric type and you need to convert it to a string type (ie, sentences), this doesn't make sense. However, if your data from the numeric type, then you can convert it into a nominal type, in WEKA, using the "weka.filters.unsupervised.attribute.NumericToNominal" filter
(http://weka.sourceforge.net/doc.dev/weka/filters/unsupervised/attribute/NumericToNominal.html).
Dr. Samer Sarsam
Hello,
I am sorry for the delay. There is no function which encodes or decodes a rule. But, weka facilitate these treatments.
I send to you my used format of rule and the code. You can see my article available on RG.
If you need other explanations, I can help you.
Good luck.
- Badges
- Science topic
- Similar topics
- Correction