Reputation: 21
| label1 | label2 | label3 | label4 | feature1 | feature2 |
| 1 | 0 | 1 | 0 | 1 | 1 |
| 0 | 1 | 1 | 0 | 0 | 0 |
I'm now having a pyspark dataframe looks like this: there are 20 labels in total and hundreds of feature columns, and all values of these columns only are 0 or 1.
I'm trying to convert all 20 labels into one column which pyspark machine learning can accept, and I'm thinking of using multilable classfication but it seems not working out. All feature columns have been transfered to sparse vector.
Is it right I can only pass single column with categorical/numerical format of values as target into spark machine learning?
I can only use pyspark dataframe to run multilable classfication/multiclass due to some restrictions. Thank you for the help!
Trying to pass multiple label columns into spark machine learning model.
Upvotes: 1
Views: 151