Currently I create a synthetic dataset using the
python sdv package and an existing dataset as below:
import pandas as pd import sdv.tabular as sd_tab file_path = 'E:/csv_files/vee_bg_data.csv' df = pd.read_csv(file_path) model = sd_tab.GaussianCopula() model.fit(df) sample = model.sample(200) print(sample.head(50).to_string())
The generated dataset has the same format properties and statistics as the provided dataset.
However, I would like to generate a dataset, where the properties of ONLY some features are different from the provided dataset. How can I do this?
A python-based solution is most preferrable, but any technique is fine.