For multi-targets, SMOTE is not recommended even if there is a bias

[mariko-sugawara]

Describe the bug
At _get_target_imbalance_score(), if the target column has multiclass, the imbalance score is calculated as 0 and SMOTE is not recommended for preprocess.

core/sapientml_core/meta_features.py

Line 1012 in f0d33b1

def _get_target_imbalance_score(Y):

To Reproduce
Steps to reproduce the behavior:

1. Show your code calling generate_code().

script

# Paste your code here. The following is an example.
from sapientml import SapientMLGenerator
sml = SapientMLGenerator()
sml.generate_code('your arguments')

2. Attach the datasets or dataframes input to generate_code() if possible.
3. Show the generated code such as 1_default.py when it was generated.

generated code

# Paste the generated code here.

4. Show the messages of SapientML and/or generated code.

Expected behavior
A clear and concise description of what you expected to happen.

Environment (please complete the following information):

• OS: [e.g. Ubuntu 20.04]
• Docker Version (if applicable): [Docker version 20.10.17, build 100c701]
• Python Version: [e.g. 3.9.12]
• SapientML Version: 0.5.4

Additional context

• For the following code line 1020, fix the condition to vc.shape[0] > 10(if follow the comment) or delete it.

        # if there are more than 10 categories, probably it is a regression problem
        if vc.shape[0] > 2:
            return 0

• In addition, offline learning is required.