Machine Learning Portfolio Simulating and Optimizing Apparatuses, Methods and Systems

Abstract

The Machine Learning Portfolio Simulating and Optimizing Apparatuses, Methods and Systems (“MLPO”) transforms machine learning simulation request, decision tree ensembles training request, expected returns calculation request, portfolio construction request, predefined scenario construction request, portfolio returns visualization request inputs via MLPO components into machine learning simulation response, decision tree ensembles training response, expected returns calculation response, portfolio construction response, predefined scenario construction response, portfolio returns visualization response outputs. An asset return metrics calculation request datastructure is obtained. The number of sessions to utilize for calculating asset return metrics data is determined. An assets range for a session is determined and asset database table records from the assets range are processed in batches using an assets batch database table, a factor simulations batch database table, and a factor exposures batch database table. Expected returns for each batch of the assets range are calculated using the batch tables via a parallel SQL query.

What is claimed is: 
     
         1 . A database calculation engine apparatus, comprising:
 a memory;   a component collection in the memory;   a processor disposed in communication with the memory and configured to issue a plurality of processor-executable instructions from the component collection, the processor-executable instructions structured as:
 obtain, via at least one processor, an asset return metrics calculation request datastructure, the asset return metrics calculation request datastructure structured to specify a set of assets and a set of simulated market scenarios, each simulated market scenario in the set of simulated market scenarios structured to comprise a set of simulated market factor values corresponding to a set of market factors; 
 determine, via at least one processor, a number of sessions to utilize for calculating asset return metrics data; 
 determine, via at least one processor, an assets range for a session based on the determined number of sessions to utilize, the assets range comprising a set of asset database table records for the set of assets to be processed by the session; 
 create, via at least one processor, an assets batch database table, the assets batch database table structured to comprise a set of asset database table records of a specified batch size from the assets range for the session; 
 create, via at least one processor, a factor simulations batch database table, the factor simulations batch database table structured to comprise a set of simulated market factor return values for the set of asset database table records in the assets batch database table; 
 create, via at least one processor, a factor exposures batch database table, the factor exposures batch database table structured to comprise a set of factor exposure database table records matching the set of asset database table records in the assets batch database table; and 
 calculate, via at least one processor, via a parallel SQL query, expected returns for the set of asset database table records in the assets batch database table, using the factor simulations batch database table and the factor exposures batch database table. 
   
     
     
         2 . The apparatus of  claim 1 , further, comprising:
 the set of simulated market factor values for a simulated market scenario is configured as generated using a set of deep learning neural networks.   
     
     
         3 . The apparatus of  claim 1 , further, comprising:
 the set of simulated market factor values for a simulated market scenario is configured as generated using a set of multi-variate mixture datastructures.   
     
     
         4 . The apparatus of  claim 1 , further, comprising:
 the processor-executable instructions structured as:
 filter, via at least one processor, asset database table records associated with the set of assets, based on available factor exposure database table records, using a SQL statement. 
   
     
     
         5 . The apparatus of  claim 4 , further, comprising:
 the processor-executable instructions structured as:
 filter, via at least one processor, simulated market scenario database table records associated with the set of simulated market scenarios, based on a subset of market factors from the set of market factors to which the filtered asset database table records have exposure, using a SQL statement. 
   
     
     
         6 . The apparatus of  claim 1 , further, comprising:
 the number of sessions to utilize for calculating asset return metrics data is determined based on the number of available server processors and a specified degree of parallelism per session.   
     
     
         7 . The apparatus of  claim 1 , further, comprising:
 the assets batch database table, the factor simulations batch database table, and the factor exposures batch database table are temporary database tables.   
     
     
         8 . The apparatus of  claim 4 , further, comprising:
 the processor-executable instructions structured as:
 filter, via at least one processor, at least one of call schedule database table records and put schedule database table records based on the filtered asset database table records. 
   
     
     
         9 . The apparatus of  claim 8 , further, comprising:
 the processor-executable instructions structured as:
 adjust, via at least one processor, via a parallel SQL query, the calculated expected returns for the set of asset database table records in the assets batch database table, based on at least one of the filtered call schedule database table records and the filtered put schedule database table records. 
   
     
     
         10 . The apparatus of  claim 1 , further, comprising:
 the processor-executable instructions structured as:
 transpose, via at least one processor, the calculated expected returns for the set of asset database table records in the assets batch database table into a wide array format; and 
 write, via at least one processor, via a parallel SQL query, the transposed expected returns to an asset simulation wide table. 
   
     
     
         11 . The apparatus of  claim 10 , further, comprising:
 the asset simulation wide table formatted to facilitate efficient calculation of portfolio return metrics; and   the asset simulation wide table structured to be written to in parallel by query server processes from a plurality of utilized sessions.   
     
     
         12 . The apparatus of  claim 1 , further, comprising:
 the processor-executable instructions structured as:
 calculate, via at least one processor, via a parallel SQL query, an asset return metric based on the calculated expected returns for the set of asset database table records in the assets batch database table; 
 transpose, via at least one processor, the calculated asset return metric for the set of asset database table records in the assets batch database table into a wide array format; and 
 write, via at least one processor, via a parallel SQL query, the transposed asset return metric to an asset simulation wide table. 
   
     
     
         13 . The apparatus of  claim 1 , further, comprising:
 the processor-executable instructions structured as:
 write, via at least one processor, via a parallel SQL query, the calculated expected returns to an asset measure table. 
   
     
     
         14 . The apparatus of  claim 1 , further, comprising:
 the processor-executable instructions structured as:
 calculate, via at least one processor, via a parallel SQL query, an asset return metric based on the calculated expected returns for the set of asset database table records in the assets batch database table; and 
 write, via at least one processor, via a parallel SQL query, the calculated asset return metric to an asset measure table. 
   
     
     
         15 . The apparatus of  claim 14 , further, comprising:
 the asset measure table formatted to facilitate efficient calculation of security return metrics; and   the asset measure table structured to be written to in parallel by query server processes from a plurality of utilized sessions.   
     
     
         16 . The apparatus of  claim 1 , further, comprising:
 the set of simulated market factor values for a simulated market scenario is configured as generated using a set of deep learning neural networks and using a set of multi-variate mixture datastructures.   
     
     
         17 . The apparatus of  claim 1 , further, comprising:
 each expected return configured as calculated for an asset during a simulated market scenario using:
 the respective asset's conditional Beta during the respective simulated market scenario, determined using a set of decision tree ensembles, trained to estimate conditional Beta of the respective asset, based on a first subset of the set of simulated market factor values, and 
 the respective asset's conditional default probability during the respective simulated market scenario, determined using a set of decision tree ensembles, trained to estimate conditional default probability of the respective asset, based on a second subset of the set of simulated market factor values. 
   
     
     
         18 . A database calculation engine processor-readable, non-transient medium, comprising processor-executable instructions structured as:
 obtain, via at least one processor, an asset return metrics calculation request datastructure, the asset return metrics calculation request datastructure structured to specify a set of assets and a set of simulated market scenarios, each simulated market scenario in the set of simulated market scenarios structured to comprise a set of simulated market factor values corresponding to a set of market factors;   determine, via at least one processor, a number of sessions to utilize for calculating asset return metrics data;   determine, via at least one processor, an assets range for a session based on the determined number of sessions to utilize, the assets range comprising a set of asset database table records for the set of assets to be processed by the session;   create, via at least one processor, an assets batch database table, the assets batch database table structured to comprise a set of asset database table records of a specified batch size from the assets range for the session;   create, via at least one processor, a factor simulations batch database table, the factor simulations batch database table structured to comprise a set of simulated market factor return values for the set of asset database table records in the assets batch database table;   create, via at least one processor, a factor exposures batch database table, the factor exposures batch database table structured to comprise a set of factor exposure database table records matching the set of asset database table records in the assets batch database table; and   calculate, via at least one processor, via a parallel SQL query, expected returns for the set of asset database table records in the assets batch database table, using the factor simulations batch database table and the factor exposures batch database table.   
     
     
         19 . A database calculation engine processor-implemented system, comprising:
 means to process processor-executable instructions;   means to issue processor-issuable instructions from a processor-executable component collection via the means to process processor-executable instructions, the processor-issuable instructions structured as:
 obtain, via at least one processor, an asset return metrics calculation request datastructure, the asset return metrics calculation request datastructure structured to specify a set of assets and a set of simulated market scenarios, each simulated market scenario in the set of simulated market scenarios structured to comprise a set of simulated market factor values corresponding to a set of market factors; 
 determine, via at least one processor, a number of sessions to utilize for calculating asset return metrics data; 
 determine, via at least one processor, an assets range for a session based on the determined number of sessions to utilize, the assets range comprising a set of asset database table records for the set of assets to be processed by the session; 
 create, via at least one processor, an assets batch database table, the assets batch database table structured to comprise a set of asset database table records of a specified batch size from the assets range for the session; 
 create, via at least one processor, a factor simulations batch database table, the factor simulations batch database table structured to comprise a set of simulated market factor return values for the set of asset database table records in the assets batch database table; 
 create, via at least one processor, a factor exposures batch database table, the factor exposures batch database table structured to comprise a set of factor exposure database table records matching the set of asset database table records in the assets batch database table; and 
 calculate, via at least one processor, via a parallel SQL query, expected returns for the set of asset database table records in the assets batch database table, using the factor simulations batch database table and the factor exposures batch database table. 
   
     
     
         20 . A database calculation engine processor-implemented process, comprising executing processor-executable instructions to:
 obtain, via at least one processor, an asset return metrics calculation request datastructure, the asset return metrics calculation request datastructure structured to specify a set of assets and a set of simulated market scenarios, each simulated market scenario in the set of simulated market scenarios structured to comprise a set of simulated market factor values corresponding to a set of market factors;   determine, via at least one processor, a number of sessions to utilize for calculating asset return metrics data;   determine, via at least one processor, an assets range for a session based on the determined number of sessions to utilize, the assets range comprising a set of asset database table records for the set of assets to be processed by the session;   create, via at least one processor, an assets batch database table, the assets batch database table structured to comprise a set of asset database table records of a specified batch size from the assets range for the session;   create, via at least one processor, a factor simulations batch database table, the factor simulations batch database table structured to comprise a set of simulated market factor return values for the set of asset database table records in the assets batch database table;   create, via at least one processor, a factor exposures batch database table, the factor exposures batch database table structured to comprise a set of factor exposure database table records matching the set of asset database table records in the assets batch database table; and   calculate, via at least one processor, via a parallel SQL query, expected returns for the set of asset database table records in the assets batch database table, using the factor simulations batch database table and the factor exposures batch database table.