`ValueError: y 中最少填充的类只有 1 个成员，这在 PyCaret 中太少了`

Question

I have a problem, working with PyCaret. Previously I did not have any problems.

But it started when I oversampled data and saved it, using pandas and this question .

The file is here .

Then I read the file in a separate notebook.

import pycaret
from pycaret.utils import version
from pycaret.regression import *
from pycaret.classification import *

# Read clean data
starbucks_days = pd.read_csv('days_smote.csv')

# Drop a column
starbucks_days = starbucks_days.drop(['Unnamed: 0'], axis = 1)
starbucks_days = starbucks_days.drop(['transaction', 'offer_viewed', 'offer_received', 'offer_completed'], axis = 1)
starbucks_days = starbucks_days.drop(['label'], axis = 1)

The I start to use PyCaret

# Initialize Setup
starbucks_days1 = setup(starbucks_days, target = 'time_completed_viewed', session_id = 123, log_experiment = True, experiment_name = 'days1')

But get an error

ValueError: The least populated class in y has only 1 member, which is too few. The minimum number of groups for any class cannot be less than 2.

This GitHub issue gives some hints

I check some parameters

type(starbucks_days)
pandas.core.frame.DataFrame

starbucks_days['time_completed_viewed'].value_counts()
6.000000      1682
12.000000     1503
18.000000     1318
24.000000     1212
174.000000    1068
          ... 
444.107530       1
226.213225       1
411.947513       1
236.001744       1
394.722944       1
Name: time_completed_viewed, Length: 3572, dtype: int64

Any tips what am I missing? As I said, PyCaret works just fine with simple csv files, which were not oversampled.

Answer 1

In your imports, you have imported classification after importing regression that has overwritten the module in the environment.

This seems like a regression problem (continuous value). You don't need to import classification .

Get rid of this line from your code and it should work fine:

from pycaret.classification import *