Amazon Technologies, Inc.

Fair Isaac Corporation

Zuohua Zhang

Joseph P Milana

Gregory Gancarz

Linyu Yang

David Griegel

Bo Zhang

Danfeng Li

A determination is made at a machine learning service that a training data set comprising a majority category of observation records and one or more minority categories of observation records meets a criterion for automated sampling. A sampling ratio to be used for a particular category of the majority category and the one or more minority categories is identified. A selected sampling methodology is applied to the particular category to obtain a sample in accordance with the sampling ratio. A particular machine learning model is trained using a result of applying at least the selected sampling methodology on the particular category.

Category-based sampling of machine learning data

A merchant profile builder (MPB) is provided. The complete merchant profiling process is split into three pieces: Data Preprocessing, Weekly Summary Calculation, and Merchant Profile Calculation. In the Data Preprocessing stage, authorization data and daily fraud data are processed on a weekly basis to yield: an authorization extract file and fraud authorization data, each of which are sorted by a merchant key. In the Weekly Summary Calculation stage, the files output by the Data Preprocessing stage are used as input to compile basic statistics for each merchant based only on data for one week. There are separate weekly summaries for authorizations and fraud authorizations. In the Merchant Profile Calculation stage, the weekly summaries from the 16 most recent weeks are combined together and averages are calculated and output a set of merchant profiles and a set of default profiles.

Method and apparatus for a merchant profile builder

Growing community of inventors

Bellevue, WA, United States of America

Zuohua Zhang