The ability to use AI and ML at scale is enabling novel solutions to complex problems once difficult or impossible to solve. One such problem is the identification of statistically significant relationships among large numbers of traditional and alternative time series data sets. Discovering meaningful relationships between security prices and such data as interest rates, news stories, Securities and Exchange Commission Filings (SEC) filings, analyst reports, economic events, corporate events, and alternative events such as weather long has been the foundation for generating meaningful trading and investment decisions.
Synechron’s Artificial Intelligence for Signal Discovery is a powerful Data Science platform that ingests large volumes of structured and unstructured data and uses the latest advancements in parallel computing to provide analysts, traders, portfolio managers, and researchers with a fast, intuitive, and scalable platform for rapidly analyzing massive collections of data and identifying meaningful correlations and Granger Causal relationships among time series. It offers a rich collection of data and a library of ML and Deep Learning algorithms with the ability to add custom data and algorithms via an intuitive user interface (UI).
HOW WE HELP
Artificial Intelligence for Signal Discovery uses TensorFlow on Spark for Natural Language Processing (NLP) and the creation of convolutional neural networks (CNN’s) for classification. Its highly-scalable, high-performance infrastructure is powered by Spark, Hadoop, Yarn, and Oozie on an Azure Linux instance with Tesla K80 Nvidia chips capable of querying billions of rows per seconds across a wide range of data sources. It can be deployed locally or in the cloud.
Discovering meaningful relationships
Synechron’s Artificial Intelligence for Signal Discovery:
Can be deployed locally or in the cloud
Is highly-scalable and high-performance
Is capable of querying billions of rows per seconds across a wide range of data sources
Utilizes the latest advancements in parallel computing, AI and ML
Provides a fast, intuitive, and scalable platform
Rapidly-analyzes massive collections of data
Identifies meaningful correlations and Granger Causal relationships among time series
Transforms unstructured data into meaningful time series data
Offers a rich collection of data and a library of ML & deep learning algorithms
Has the ability to add custom data and algorithms easily
Is user-friendly with an intuitive user interface
Provides rich visuals that makes it accessible to the broadest set of users
Synechron’s Artificial Intelligence for Signal Discovery has the following data available for analysis:
Five years of U.S. historical equity prices at 30-second intervals across multiple exchanges
Historical weather data for various cities
Google GDELT data – global event data
Business event data derived from news feeds
Tagged news feeds
Treasury and consumer finance data from The Federal Reserve
U.S. economic data from FRED – The St. Louis Federal Reserve
Historical/current mortgage loan-level data from FreddieMac/FannieMae
Detailed mortgage loan data and the current loan status
Bond prices, trades
/ /THE WORK WE DO
Syn-AI is a high-compute technology architecture that power’s Synechron’s Data Science Accelerators. It uses industry standard technologies such as Spark and Hadoop to create a highly-flexible data ingestion architecture.
A statistically significant Granger Causality test between BAC’s Vader scores and FISV’s stock return indicates a potentially causal relationship between news about BAC and the price of FISV stock, especially if neither FISV nor BAC has significant Granger Causal relationships with AAPL.
Synechron’s Artificial Intelligence for Signal Discovery is designed to identify statistically significant relationships. To learn more about Synechron’s Artificial Intelligence for Signal Discovery Accelerator for financial services and the work we’re doing email us at firstname.lastname@example.org