Capital One Services, LLC

Trading Technologies International, Inc.

Carolyn Phillips

David I Widerhorn

Paul R Giedraitis

Venkataseshagiri Chintala

Melanie Rubino

Melanie A Rubino

Paul Giedraitis

Various techniques are described for using machine-learning artificial intelligence to improve how trading data can be processed to detect improper trading behaviors such as trade spoofing. In an example embodiment, semi-supervised machine learning is applied to positively labeled and unlabeled training data to develop a classification model that distinguishes between trading behavior likely to qualify as trade spoofing and trading behavior not likely to qualify as trade spoofing. Also, clustering techniques can be employed to segment larger sets of training data and trading data into bursts of trading activities that are to be assessed for potential trade spoofing status.

Applied artificial intelligence technology for processing trade data to detect patterns indicative of potential trade spoofing

Methods and systems of similarity searching encrypted data strings are disclosed. An exemplary method can include receiving data strings, obtaining a set of reference strings, determining edit distances between each data string and the reference strings, converting each set of edit distances into a document of tokens. A method may further include encrypting the data strings, associating each of the documents with a corresponding data string, and storing the data strings and the associated documents in a memory. A method may continue by receiving a search request, determining a search set of edit distances between the search request and the reference strings, converting the search set of edit distances into a document, comparing the search document with the documents stored in memory to determine which documents are above a similarity threshold compared to the search document, and returning the data strings associated with documents above the similarity threshold.

Methods and systems for similarity searching encrypted data strings

A non-transitory computer-readable storage medium is provided to store computer-readable program code to receive an unmerged record set, comprising a first plurality of data records, to generate record-pairs from the first plurality of data records, based upon a set of transitive deterministic matching criteria, apply a set of non-transitive matching rules to the record-pairs, perform a partitioning operation on the record-pairs, using a plurality of independent grouping operations, wherein a plurality of matched record groups are generated. The computer-readable program code may determine a set of maximal connected components from the plurality of matched record groups, perform a merge operation on the set of maximal connected components to generate a set of merged records, the set of merged records comprising a second plurality of data records, less than the first plurality of data records, and send the merged records for storage in a non-transitory computer readable storage medium.

System and techniques for data record merging

Growing community of inventors

Chicago, IL, United States of America

Carolyn Phillips