Abstract: A server computer system may determine a set of data related to a set of jobs processed by a computing platform on behalf of subscribers to the computing platform as part of an attribute analysis to facilitate the detection of fraudulent activity on the computing platform. The set of data may include, for each job, a set of attribute-value pairs corresponding to a set of attributes utilized to parameterize each job in the set of jobs. An ML model may generate a ranked list of the set of attribute-value pairs based on predictive utility of each attribute-value pair for identifying jobs involving fraudulent activity. The server computer system may determine a set of statistics for a threshold number of top attribute-value pairs in the ranked list in. Further, the server computer system may present the set of statistics for the threshold number of top attribute-value pairs via a graphical user interface.

Abstract: In an example embodiment, a solution is provided wherein a single machine learning model is trained and used to detect fraud across multiple payment types. More particularly, the concept of a payment scoring event (PSE) is introduced. A PSE is a data structure that stores multiple pieces of information about a transaction (or potential transaction). A mapping is then maintained between each payment type to be supported and the PSE structure. Each of these transaction types may have its own mapping indicating which fields in the PSR its own fields map to. The use of these mappings allows for additional payment methods to be introduced and supported at any time, necessitating only the creation of a mapping for an additional payment method.

Abstract: Embodiments provide resource-efficient and scalable techniques to process transaction-related data and make the transaction-related data available in a uniform and accessible manner. In one embodiment, a transaction event data pipeline may be utilized to collect events, process the events, such as by removing irrelevant or duplicative data. Further, enriched events may be generated based on the event processing. For example, enriched events may combine a plurality of events from one or more sources into a single event with a predetermined format. In several embodiments, generation of enriched events may facilitate scalability, such as by reducing the state size of various event processing components.

Abstract: This disclosure describes targeted heuristic rule generation tools for fraudulent activity. More specifically, embodiments are directed to a server system for implementing a transaction processing rule (TPR) generator that facilitates generation of transaction processing rules. In many embodiments, these transaction processing rules may be directed to blocking (or allowing) transactions in scenarios that are generally uncommon, but disproportionately affect some entities (e.g., merchants). For example, some merchants may be particularly vulnerable to certain types of fraud that a majority of merchants are not vulnerable to, such as repetitive order and refund fraud schemes. Embodiments may include various components that operate to assist a user (e.g., a merchant) in creating and implementing transaction processing rules tailored to unique or uncommon scenarios they may face.

Abstract: In an example embodiment, a machine learning training pipeline is introduced that continuously monitors and processes training data having multiple transaction types using a sliding window, adding labels as they are available for the various different types of transactions in the training data. The processed training data, with the appropriate labels added, can then be utilized by any machine learning model that is being onboarded using the pipeline, without any specialized setup being necessary. Further, even if new data is added to the pipeline to aid in the training of a new model (such as data regarding a new payment type), this new data can be processed quickly and added to the existing data without requiring specialized processes by the entity requesting the new machine learning model. This allows the actual training of the new machine learning model to be accomplished very quickly, and deployment to be accomplished even faster.

Abstract: In an example embodiment, a solution is provided wherein a single machine learning model is trained and used to detect fraud across multiple payment types. More particularly, the concept of a payment scoring event (PSE) is introduced. A PSE is a data structure that stores multiple pieces of information about a transaction (or potential transaction). A mapping is then maintained between each payment type to be supported and the PSE structure. Each of these transaction types may have its own mapping indicating which fields in the PSR its own fields map to. The use of these mappings allows for additional payment methods to be introduced and supported at any time, necessitating only the creation of a mapping for an additional payment method.