公开(公告)号：US20230229941A1

公开(公告)日：2023-07-20

申请号：US18189784

申请日：2023-03-24

Applicant: Microsoft Technology Licensing, LLC

Inventor： Edmund Chi Man Tse ,  Brett Owens Simons ,  Sandeep Repaka ,  Yatpang Cheung

IPC: G06N5/025 ,  G06N20/20 ,  G06F18/214 ,  G06F18/2113 ,  G06F18/24 ,  G06N5/01

CPC classification number: G06N5/025 ,  G06N20/20 ,  G06F18/214 ,  G06F18/2113 ,  G06F18/24765 ,  G06N5/01

Abstract: Rule induction is used to produce human readable descriptions of patterns within a dataset. A rule induction algorithm or classifier is a type supervised machine learning classification algorithm. A rule induction classifier is trained, which involves using labelled examples in the dataset to produce a set of rules. Rather than using the rules/classifier to make predictions on new unlabeled samples, the training of the rule induction model outputs human-readable descriptions of patterns (rules) within the dataset that gave rise to the rules (rather than using the rules to predict new unlabeled samples). Parameters of the rule induction algorithm are tuned to favor simple and understandable rules, instead of only tuning for predictive accuracy. The learned set of rules are outputted during the training process in a human-friendly format.

公开(公告)号：US11640543B2

公开(公告)日：2023-05-02

申请号：US16443859

申请日：2019-06-17

Applicant: Microsoft Technology Licensing, LLC

Inventor： Edmund Chi Man Tse ,  Brett Owens Simons ,  Sandeep Repaka ,  Yatpang Cheung

IPC: G06N5/02 ,  G06N20/20 ,  G06K9/62 ,  G06N5/00 ,  G06N5/025

Abstract: Rule induction is used to produce human readable descriptions of patterns within a dataset. A rule induction algorithm or classifier is a type supervised machine learning classification algorithm. A rule induction classifier is trained, which involves using labelled examples in the dataset to produce a set of rules. Rather than using the rules/classifier to make predictions on new unlabeled samples, the training of the rule induction model outputs human-readable descriptions of patterns (rules) within the dataset that gave rise to the rules (rather than using the rules to predict new unlabeled samples). Parameters of the rule induction algorithm are tuned to favor simple and understandable rules, instead of only tuning for predictive accuracy. The learned set of rules are outputted during the training process in a human-friendly format.