Attributes do not match

Probably you have not done the same exact preprocessing ETL in the model development set as in the new data you are trying to apply the model to.  You need to make sure you do all the same transformations, etc., otherwise your model may be trying to use attributes that don't exist or have been modified from their raw form.

I did all the exact preprocessing for both

Well, somehow RapidMiner determines that the attributes are not the same, since that is what the error is telling you.  You should do a row by row comparison of the attributes in both files and double check things like attribute names and data types (sometimes a data type will come in one file as numeric and another file as polynominal, or other similar problems).

I will explain how I did it to ensure I'm on the right road. All labels were sat as binomial before importing them. Then I use multi label classification binary relevance (0, 1) in training phase. In application phase, I provide only the texts alongside labels to let machine predict (0 or 1) for each label. As I said I follow the exact preprocessing 

Hi @Nawaf,

In order we can reproduce what you observe and thus understand what is going on, can you please share your process and data ?

Regards,

Lionel

This is a sample of the trained data
This is its main and inner process 

This is a sample of data need to be predicted

This is its main and inner process 

Hopefully you could help me to discover the error

@Nawaf,

To be honest, the pictures you shared are too small to be read, thus these pictures are unusable : 
Can you please share :  
 - your RM process (.rmp file) via File -> Export Process
 - your different data files (a priori Excel files)

Thank you for your understanding,

Regards,

Lionel

sorry for the confusion, how to remove unrelated file from previous post

click on the wheel of your psot and click on "Edit" and delete what you want to delete

Regards,

Lionel  : 

PS : I'm working on your process, I think that you did not connect the wordlist got from your training process (that you Store) to the 
word input port of your Process documents from data of your "prediction process". 
This way the word attributes are strictly the same in your "training process" and in your "prediction process" which is a mandatory condition for your "prediction process" to work as explained by @Telcontar120...

It doesn't allow me to delete attachments

So wordlist  is the problem?

@Nawaf

Please try this "prediction process" (in attached file) with your real data.
You will see that I connected the wordlist that you store in your "training process" to the wordlist input port of your Process Documents from Data of your "prediction process".

Please try it and tell me if the prediction process is working fine now  ….

Regards,


Lionel 

PS : Note that this process will still raise an error because you have only missing values on your testing excel file...

Thanks a lot!
Is this model for training or unseen data prediction?

@Nawaf,

The model I shared is for unseen data prediction (to be executed AFTER execution of the training process)

Regards,

Lionel

Why you put multi-label modeling operators again since I already stored it in training phases

I'm not sure if you check both models I did. One for training that worlist and model stored and other one for prediction loaded the stored worlist and model. I attached them 

These

lionelderkrikor 
Sorry for the confusion but if you check both files, you can see I divided this model into two phases and did the wordlist connection in prediction file

@Nawaf ,

I think I understood:  it is linked to the fact that you did not give a "special role" to the attributes you want to predict (the labels) : 
To be sure to understand you want to predict all your attributes (surveillance, compulsion etc. ) based on your "Text" attribute" ,right ?

Regards,

Lionel

@lionelderkrikor
I did but gave the same error  

@Nawaf,

You can find in attached file a "training process" which gives a special role to all attributes you want to predict (with a Loop attributes operator).
Please run this "training" process first and then run your "Prediction.rmp" process. In my case, this last process is working fine.

Hope this helps,

Regards,

Lionel 

You are the best, it finally works.