Granted Patent B2

US 12,725,053 · App. 17/968,215 · Granted Sep 1, 2026

Inventors: Keon Shik Kim (Cambridge, MA); Cole Witaszek (Brooklyn, NY); Abram Turner (Brooklyn, NY)

Assignee: Nasdaq, Inc.

G06N5/022

Abstract

Claims (55)

1 . A computer system to optimize, using machine learning and predictive control, allocation of computer and data communication resources for listing option contracts on an electronic trading platform, comprising:

a transceiver configured to receive over a data communications network different input data sets from one or more trading source computers communicating with the data communications network, where an input data set includes trading symbol data objects, each trading symbol data object corresponding to an underlying asset having an associated trading symbol and including associated trading symbol data object attributes corresponding to different options contracts for the underlying asset having the associated trading symbol;

a processing system that includes at least one hardware processor, the processing system configured to:

(a) process the input data sets using a predictive machine learning model to predict for a predetermined future time period a set of trading symbol data objects, each trading symbol data object having predicted future trading symbol data object attribute values including a predicted future volume of option contract trading activity for the trading symbol for the predetermined future time period;

(b) sort the set of trading symbol data objects based on the predicted future trading symbol data object attribute values to generate a current, ranked list of trading symbol data objects with predicted future trading symbol data object attribute values including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period;

(c) modify the current, ranked list of trading symbol data objects to account for a prior ranking of trading symbol data objects to generate a modified, ranked list of trading symbol data objects with predicted future trading symbol data object attribute values including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period;

(d) replace a subset of lower ranked trading symbol data objects from the prior ranking of trading symbol data objects with a subset of higher ranked trading symbol data objects from the modified, ranked list of trading symbol data objects with predicted future trading symbol data object attribute values to generate a new ranking of trading symbol data objects with predicted future trading symbol data object attribute values including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period;

(e) calculate one or more performance metrics for trading symbol data objects in the new ranking of trading symbol data objects relative to trading symbol option listings for the trading symbol actually traded over the predetermined future time period for each of the trading symbol data objects in the new ranking of trading symbol data objects;

(f) retrain the predictive machine learning model based on the one or more performance metrics indicating a need to improve performance of the predictive machine learning model;

(g) allocate computer system resources for the new ranking of trading symbol data symbols for the predetermined time period; and

(h) generate and output a display or a file including the new ranking of trading symbol data objects.

2 . The computer system in claim 1 , wherein the processing system is configured to iterate (a)-(f) as a part of retraining the predictive machine learning model.

3 . The computer system in claim 1 , wherein the processing system is configured to:

process the input data sets to generate formatted input data for the predictive machine learning model.

4 . The computer system in claim 1 , wherein the processing system is configured to sort the set of trading symbol data objects by highest predicted future trading symbol data object attribute value corresponding to predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period.

5 . The computer system in claim 1 , wherein the processing system is configured to:

process the input data sets using the predictive machine learning model to predict for the predetermined future time period a first list of trading symbol data objects with predicted future trading symbol data object attributes including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period and a second list of trading symbol data objects with predicted future trading symbol data object attributes including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period, and

perform (b)-(d) using the first list of trading symbol data objects and the second list of trading symbol data objects.

6 . The computer system in claim 1 , wherein the processing system is configured to divide the current, ranked list of predicted future values for the trading symbol data objects into a first set of trading symbol data objects to be added to the new ranking of trading symbol data objects and a second set of trading symbol data objects to be removed from the new ranking of trading symbol data objects.

7 . The computer system in claim 1 , wherein the trading symbol data object attributes include a date, a strike price, and quantity or amount associated with one of the trading symbol data objects.

8 . The computer system in claim 1 , wherein the processing system is configured to replace the subset of lower ranked trading symbol data objects from the prior ranking of trading symbol data objects with the subset of higher ranked trading symbol data objects from the modified ranked list of predicted selection values based on one or both of:

replacement rules; and

replacement parameters.

9 . The computer system in claim 8 , wherein the processing system is configured to adjust one or both of the replacement rules and the replacement parameters based on the one or more performance metrics.

10 . The computer system in claim 8 , wherein the processing system is configured to modify the current ranked list of predicted values to account for an impact associated with a trading symbol data object's prior status or other extenuating conditions to generate a modified, ranked list of predicted values.

11 . A method to optimize allocation of computer and data communication resources for listing option contracts on an electronic trading platform, comprising:

storing in at least one memory input data sets from one or more trading source computers communicating with the data communications network, where each of the input data sets includes trading symbol data objects, each trading symbol data object corresponding to an underlying asset having an associated trading symbol and including associated trading symbol data object attributes corresponding to different options contracts for the underlying asset having the associated trading symbol;

at a processing system that includes at least one hardware processor:

(a) process the input data sets using a predictive machine learning model to predict for a predetermined future time period a set of trading symbol data objects, each trading symbol data object having predicted future trading symbol data object attribute values including a predicted future volume of option contract trading activity for the trading symbol for the predetermined future time period;

(b) sort the set of trading symbol data objects based on the predicted future trading symbol data object attribute values to generate a current, ranked list of trading symbol data objects with predicted future trading symbol data object attribute values including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period;

(c) modify the current, ranked list of trading symbol data objects to account for a prior ranking of trading symbol data objects to generate a modified, ranked list of trading symbol data objects with predicted future trading symbol data object attribute values including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period;

(d) replace a subset of lower ranked trading symbol data objects from the prior ranking of trading symbol data objects with a subset of higher ranked trading symbol data objects from the modified, ranked list of trading symbol data objects with predicted future trading symbol data object attribute values to generate a new ranking of trading symbol data objects with predicted future trading symbol data object attribute values including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period;

(e) calculate one or more performance metrics for trading symbol data objects in the new ranking of trading symbol data objects relative to trading symbol option listings for the trading symbol actually traded over the predetermined future time period for each of the trading symbol data objects in the new ranking of trading symbol data objects;

(f) retrain the predictive machine learning model based on the one or more performance metrics indicating a need to improve performance of the predictive machine learning model;

(g) allocate computer system resources for the new ranking of trading symbol data symbols for the predetermined time period; and

(h) generate and output a display or a file including the new ranking of trading symbol data objects.

12 . The method in claim 11 , further comprising iterating (a)-(f) as a part of retraining the predictive machine learning model.

13 . The method in claim 11 , further comprising:

processing the input data sets using the predictive machine learning model to predict for the predetermined future time period a first list of trading symbol data objects with predicted future trading symbol data object attributes including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period and a second list of trading symbol data objects with predicted future trading symbol data object attributes including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period, and

performing (b)-(d) using the first list of trading symbol data objects and the second list of trading symbol data objects.

14 . The method in claim 11 , further comprising replacing the subset of lower ranked trading symbol data objects from the prior ranking of trading symbol data objects with the subset of higher ranked trading symbol data objects from the modified ranked list of predicted selection values based on one or both of:

replacement rules,

replacement parameters.

15 . The method in claim 11 , further comprising adjusting one or both of the replacement rules and the replacement parameters based on the one or more performance metrics.

16 . The method in claim 11 , further comprising modifying the current ranked list of predicted values to account for an impact associated with a trading symbol data object's prior status or other extenuating conditions to generate a modified ranked list of predicted values.

17 . A non-transitory, computer-readable medium encoded with instructions that, when executed by at least one hardware processor, cause the at least one hardware processor to perform operations to optimize, using machine learning and predictive control, allocation of computer and data communication resources for listing option contracts on an electronic trading platform, comprising:

(a) receiving input data sets including trading symbol data objects, received over a data communications network from one or more trading source computers communicating with the data communications network, each trading symbol data object corresponding to an underlying asset having an associated trading symbol and including associated trading symbol data object attributes corresponding to different options contracts for the underlying asset having the associated trading symbol;

(b) processing the input data sets using a predictive machine learning model to predict for a predetermined future time period a set of trading symbol data objects, each trading symbol data object having predicted future trading symbol data object attribute values including a predicted future volume of option contract trading activity for the trading symbol for the predetermined future time period;

(c) ranking the set of trading symbol data objects based on the predicted future trading symbol data object attribute values to generate a current, ranked list of trading symbol data objects with predicted future trading symbol data object attribute values including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period;

(d) modifying the current, ranked list of trading symbol data objects to account for a prior ranking of trading symbol data objects to generate a modified, ranked list of trading symbol data objects with predicted future trading symbol data object attribute values including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period;

(e) replacing a subset of lower ranked trading symbol data objects from the prior ranking of trading symbol data objects with a subset of higher ranked trading symbol data objects from the modified, ranked list of trading symbol data objects with predicted future trading symbol data object attribute values to generate a new ranking of trading symbol data objects with predicted future trading symbol data object attribute values including predicted future volumes of option contract trading activity for each trading symbol for the predetermined future time period;

(f) calculating one or more performance metrics for trading symbol data objects in the new ranking of trading symbol data objects relative to trading symbol option listings for the trading symbol actually traded over the predetermined future time period for each of the trading symbol data objects in the new ranking of trading symbol data objects;

(g) retraining the predictive machine learning model based on the one or more performance metrics indicating a need to improve performance of the predictive machine learning model;

(h) allocate computer system resources for the new ranking of trading symbol data symbols for the predetermined time period; and

(i) generate and output a display or a file including the new ranking of trading symbol data objects.

Assignments (1)

Continuity (1)

References Cited (46)