Sift Science has filed a patent for a machine learning-based method to accelerate the generation of automated fraud or abuse detection workflows in a digital threat mitigation platform. The method involves identifying distinct digital event features indicative of fraud, deriving decisioning criteria based on these features and a data corpus, constructing a probationary workflow, and deploying it to a target fraud prevention environment. The method aims to improve the efficiency and effectiveness of fraud detection processes. GlobalData’s report on Sift Science gives a 360-degree view of the company including its patenting strategy. Buy the report here.
Smarter leaders trust GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
According to GlobalData’s company profile on Sift Science, AI-assisted threat classification was a key innovation area identified from patents. Sift Science's grant share as of September 2023 was 79%. Grant share is based on the ratio of number of grants to total number of patents.
Machine learning-based method for accelerating fraud detection workflows
A recently filed patent (Publication Number: US20230316282A1) describes a machine learning-based method for accelerating the generation of automated fraud or abuse detection workflows in a digital threat mitigation platform. The method involves identifying distinct digital event features indicative of digital fraud, deriving digital event decisioning criteria based on these features, and constructing a probationary automated fraud or abuse detection workflow. The workflow includes defining automated digital event decisioning routes and setting route execution conditions based on the decisioning criteria. The probationary workflow is then deployed to a target digital fraud prevention environment based on efficacy metrics computed for the workflow.
The method utilizes a digital event meta machine learning model trained on a digital event data corpus associated with the target subscriber. The model includes distinct digital event classification paths that classify digital events into different threat classes. The decisioning criteria are extracted from these classification paths. The automated digital event decisioning routes in the probationary workflow correspond to the classification paths and their execution conditions are based on the extracted decisioning criteria.
The digital event data corpus used for training the model includes distinct digital event meta-training data samples. Each sample corresponds to a digital event associated with the target subscriber and includes a digital threat score, feature values, and a chargeback label. The method also involves constructing feature distribution graphs for each distinct digital event feature and assessing them to derive the decisioning criteria.
The probationary automated fraud or abuse detection workflow can include routes that process digital events to completion or prevent them from being processed. The execution conditions for these routes are defined based on the extracted decisioning criteria. The method also involves simulating the workflow based on historical digital events, computing efficacy metrics, and iteratively tuning the workflow until it satisfies pre-determined performance criteria.
The patent also describes scenarios where the probationary workflow replaces an incumbent workflow that does not meet performance criteria or is automatically constructed for a new subscriber who has not deployed a sufficient number of automated fraud or abuse detection workflows.
Overall, this patent presents a machine learning-based method for accelerating the
To know more about GlobalData’s detailed insights on Sift Science, buy the report here.
Data Insights
From
The gold standard of business intelligence.
Blending expert knowledge with cutting-edge technology, GlobalData’s unrivalled proprietary data will enable you to decode what’s happening in your market. You can make better informed decisions and gain a future-proof advantage over your competitors.
GlobalData, the leading provider of industry intelligence, provided the underlying data, research, and analysis used to produce this article.
GlobalData Patent Analytics tracks bibliographic data, legal events data, point in time patent ownerships, and backward and forward citations from global patenting offices. Textual analysis and official patent classifications are used to group patents into key thematic areas and link them to specific companies across the world’s largest industries.
