Feature engineering pipeline generation for machine learning using decoupled dataset analysis and interpretation
LLE Rouesnel - US Patent 11,593,705, 2023 - Google Patents
the sets of data facts can be combined into sets of data statements and utilized by multiple
interpretation engines to, according to different processing strategies, generate different
feature engineering pipelines that specify operations for generating feature sets that can be
used for training ML models.ones of the feature engineering pipelines can be provided to
users, who may select a preferred feature engineering pipeline for use or further modify the
feature engineering pipelines according to user-knowledge for further customization.
interpretation engines to, according to different processing strategies, generate different
feature engineering pipelines that specify operations for generating feature sets that can be
used for training ML models.ones of the feature engineering pipelines can be provided to
users, who may select a preferred feature engineering pipeline for use or further modify the
feature engineering pipelines according to user-knowledge for further customization.
以上显示的是最相近的搜索结果。 查看全部搜索结果