将两个不同的 Sklearn 分类器应用于相同数据的两个不同子集

Question

I have a dataset that I need to run through a classification Pipeline. The dataset has 2 types of rows:

• described: description column POPULATED
• non-desribed: description column EMPTY

I want to apply one classifier targetting ONLY the described data, and another one for the non-described data.

I am currently doing so by separating the dataset, and then preprocessing and feeding the dataset with their corresponding classifier separately. What I want to accomplish is fitting this process into a Sklearn pipeline. It should be something like this:

classifierPipe = Pipeline([('preproc_described', DescPreprocessor),
                           ('preproc_non_described', NonDescPreprocessor),
                           ('clf_described', CLF1),
                           ('clf_described', CLF2)
                          ])

classifierPipe.fit(X_train,y_train)

I was reviewing StackingClassifier , but according to the documentation, initial estimators are applied to all the rows in the dataset.

How can I create such a pipeline where each classifier targets a specific subset of the whole dataset?

Answer 1

Why not just create two different datasets and use one classifier on each. A simple code like below should be sufficient

 import pandas as pd df = pd.read_cvs('csv_name.csv') #drop each column in the resp dataset for_clf_1 = df.drop(['described'],axis = 1) for_clf_2 = df.drop(['not described'], axis =1)