AI-Based Real-Time Prediction Engine Apparatuses, Methods and Systems

Abstract

The AI-Based Real-Time Prediction Engine Apparatuses, Methods and Systems (“AIRTPE”) transforms machine learning training input, order placement input inputs via AIRTPE components into machine learning training output, order placement output, information leakage alert outputs. An order placement datastructure associated with a security identifier is obtained. An order placement allocation for the security identifier is determined. An order placement request datastructure for a first order is sent to a server associated with a first venue. A set of trade tick data messages associated with the first venue is obtained. A set of inferred labels is determined for each obtained trade tick data message using a real-time prediction logic generated using a machine learning technique. The inferred labels of a selected inferred label type are grouped into buckets. When it is determined that the grouped inferred labels correspond to execution data generated by the first order, an information leakage alert is generated.

1 - 18 . (canceled) 
     
     
         19 . An artificial intelligence-based information leakage order generating apparatus, comprising:
 a memory;   a component collection in the memory, including:
 a trade tick monitoring component; 
   a processor disposed in communication with the memory, and configured to issue a plurality of processing instructions from the component collection stored in the memory,
 wherein the processor issues instructions from the trade tick monitoring component, stored in the memory, to:
 obtain, via at least one processor, a set of trade tick data messages, wherein the set of trade tick data messages pertains to a security identifier; 
 determine, via at least one processor, a set of inferred labels for each obtained trade tick data message using a real-time prediction logic generated using a machine learning technique; 
 compute, via at least one processor, a real-time liquidity model for the security identifier using the determined inferred labels; 
 assess, via at least one processor, counterparty landscape for the security identifier using the real-time liquidity model; 
 detect, via at least one processor, information leakage associated with the security identifier based on the assessment of the counterparty landscape; 
 determine, via at least one processor, an order placement allocation for the security identifier based on the detected information leakage; and 
 send, via at least one processor, an order placement request datastructure for a first order specified by the order placement allocation to a server associated with a first venue specified by the order placement allocation for the first order. 
 
   
     
     
         20 . A processor-readable artificial intelligence-based information leakage order generating non-transient physical medium storing processor-executable components, the components, comprising:
 a component collection stored in the medium, including:
 a trade tick monitoring component; 
 wherein the trade tick monitoring component, stored in the medium, includes processor-issuable instructions to:
 obtain, via at least one processor, a set of trade tick data messages, wherein the set of trade tick data messages pertains to a security identifier; 
 determine, via at least one processor, a set of inferred labels for each obtained trade tick data message using a real-time prediction logic generated using a machine learning technique; 
 compute, via at least one processor, a real-time liquidity model for the security identifier using the determined inferred labels; 
 assess, via at least one processor, counterparty landscape for the security identifier using the real-time liquidity model; 
 detect, via at least one processor, information leakage associated with the security identifier based on the assessment of the counterparty landscape; 
 determine, via at least one processor, an order placement allocation for the security identifier based on the detected information leakage; and 
 send, via at least one processor, an order placement request datastructure for a first order specified by the order placement allocation to a server associated with a first venue specified by the order placement allocation for the first order. 
 
   
     
     
         21 . A processor-implemented artificial intelligence-based information leakage order generating system, comprising:
 a trade tick monitoring component means, to:
 obtain, via at least one processor, a set of trade tick data messages, wherein the set of trade tick data messages pertains to a security identifier; 
 determine, via at least one processor, a set of inferred labels for each obtained trade tick data message using a real-time prediction logic generated using a machine learning technique; 
 compute, via at least one processor, a real-time liquidity model for the security identifier using the determined inferred labels; 
 assess, via at least one processor, counterparty landscape for the security identifier using the real-time liquidity model; 
 detect, via at least one processor, information leakage associated with the security identifier based on the assessment of the counterparty landscape; 
 determine, via at least one processor, an order placement allocation for the security identifier based on the detected information leakage; and 
 send, via at least one processor, an order placement request datastructure for a first order specified by the order placement allocation to a server associated with a first venue specified by the order placement allocation for the first order. 
   
     
     
         22 . A processor-implemented artificial intelligence-based information leakage order generating method, comprising:
 executing processor-implemented trade tick monitoring component instructions to:
 obtain, via at least one processor, a set of trade tick data messages, wherein the set of trade tick data messages pertains to a security identifier; 
 determine, via at least one processor, a set of inferred labels for each obtained trade tick data message using a real-time prediction logic generated using a machine learning technique; 
 compute, via at least one processor, a real-time liquidity model for the security identifier using the determined inferred labels; 
 assess, via at least one processor, counterparty landscape for the security identifier using the real-time liquidity model; 
 detect, via at least one processor, information leakage associated with the security identifier based on the assessment of the counterparty landscape; 
 determine, via at least one processor, an order placement allocation for the security identifier based on the detected information leakage; and 
 send, via at least one processor, an order placement request datastructure for a first order specified by the order placement allocation to a server associated with a first venue specified by the order placement allocation for the first order. 
   
     
     
         23 . The apparatus of  claim 19 , wherein the order placement allocation specifies a plurality of orders to place in a plurality of venues. 
     
     
         24 . The apparatus of  claim 19 , wherein the inferred label is any one of: predicted trading algorithm type, predicted client type, predicted tier, predicted state, predicted venue, predicted limit price. 
     
     
         25 . The apparatus of  claim 19 , wherein the real-time liquidity model includes specified buckets having time buckets having a specified size. 
     
     
         26 . The apparatus of  claim 25 , wherein the information leakage is an allocation information leakage alert generated based on determining that the predicted trading algorithm type corresponds to the trading algorithm type associated with the trading algorithm identifier. 
     
     
         27 . The apparatus of  claim 26 , further, comprising:
 the processor issues instructions from the trade tick monitoring component, stored in the memory, to:
 modify, via at least one processor, the order placement allocation for the security identifier using a different trading algorithm; and 
 send, via at least one processor, an order placement request datastructure for a second order specified by the modified order placement allocation. 
   
     
     
         28 . The apparatus of  claim 19 , wherein the order placement datastructure further includes a limit price, and wherein the specified buckets are price buckets. 
     
     
         29 . The apparatus of  claim 28 , wherein the information leakage alert is a limit price information leakage alert generated based on determining that the predicted limit price is within a threshold of the limit price. 
     
     
         30 . The apparatus of  claim 29 , further, comprising:
 the processor issues instructions from the trade tick monitoring component, stored in the memory, to:
 modify, via at least one processor, the order placement allocation for the security identifier to use a different limit price; and 
 send, via at least one processor, an order placement request datastructure for a second order having the different limit price. 
   
     
     
         31 . The apparatus of  claim 19 , further, comprising:
 the processor issues instructions from the trade tick monitoring component, stored in the memory, to:
 compute, via at least one processor, a real-time liquidity model for the security identifier using the determined inferred labels; 
 assess, via at least one processor, counterparty landscape for the first order using the real-time liquidity model; and 
 modify, via at least one processor, an internal state of the trading algorithm associated with the trading algorithm identifier based on the counterparty landscape. 
   
     
     
         32 . The apparatus of  claim 31 , wherein the order placement allocation for the security identifier is further determined based on the modified internal state of the trading algorithm. 
     
     
         33 . The apparatus of  claim 31 , further, comprising:
 the processor issues instructions from the trade tick monitoring component, stored in the memory, to:
 modify, via at least one processor, the order placement allocation for the security identifier based on the modified internal state of the trading algorithm; and 
 send, via at least one processor, an order placement request datastructure for a second order specified by the modified order placement allocation. 
   
     
     
         33 . The apparatus of  claim 31 , wherein the real-time liquidity model for the security identifier indicates at least one of: percentages of orders that are executed in different markets, percentages of orders that are executed using different trading algorithms. 
     
     
         34 . The apparatus of  claim 19 , wherein the machine learning technique used to generate the real-time prediction logic is one of: random forest, gradient boosting, decision tree, logistic regression. 
     
     
         35 . The apparatus of  claim 34 , further, comprising:
 a machine learning training component;   wherein the processor issues instructions from the machine learning training component, stored in the memory, to:
 retrieve, via at least one processor, historical trade tick data; 
 retrieve, via at least one processor, historical quote data; 
 augment, via at least one processor, the historical trade tick data with the historical quote data, wherein an as of join of the historical quote data and the historical trade tick data is performed; 
 retrieve, via at least one processor, historical order execution data; 
 determine, via at least one processor, venues for the historical order execution data; 
 determine, via at least one processor, time buffers for the historical order execution data, wherein each venue is associated with a separate time buffer; 
 generate, via at least one processor, labeled trade tick data, wherein an as of join of the historical order execution data and the augmented historical trade tick data is performed; 
 select, via at least one processor, a subset of the labeled trade tick data that avoids overfitting using sampling stratified over a set of buckets; and 
 train, via at least one processor, the real-time prediction logic using the machine learning technique and the selected subset of the labeled trade tick data.