Options
How to build a individualized ensemble classifier with RM 5
catchyoulater
Member Posts: 4 
Contributor I
Hi,
My task is a binomial classification question of a very large data set, including 8,900,000 rows. In considering the size of the data set, I tried to built the Bagging myself, cause the Bagging in the RM 5.0 seems not support the large data set I faced. I did this just follow these steps: I, firstly, spited the data to 100 parties then I use each party to train a different classification algorithm. After the 100 classifiers were trained, I want to find an operator to combine my 100 classifiers by weight, but I cannot to find the very operator. Is anyone finding the operator which can solve my problem? Any suggestions will be appreciated.
My task is a binomial classification question of a very large data set, including 8,900,000 rows. In considering the size of the data set, I tried to built the Bagging myself, cause the Bagging in the RM 5.0 seems not support the large data set I faced. I did this just follow these steps: I, firstly, spited the data to 100 parties then I use each party to train a different classification algorithm. After the 100 classifiers were trained, I want to find an operator to combine my 100 classifiers by weight, but I cannot to find the very operator. Is anyone finding the operator which can solve my problem? Any suggestions will be appreciated.
Tagged:
0
Answers
there's an operator called Generate Aggregation that will aggregate a number of attributes of each single example. Don't exactly know if this could be helpful in your situation? How are the outcomes of the 100 models represented in the dataset?
Another question: Which problems does the Bagging operator cause? Out of memory? How many attributes does your dataset have?
Greetings,
Sebastian