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. A portfolio construction request configured to include a set of optimization parameters is obtained. A set of simulated market scenarios is generated using neural networks. A set of expected returns for securities in the universe of securities for the set of simulated market scenarios is retrieved. Portfolio weights of securities in the universe of securities are optimized to generate a set of tradeable transactions that maximize expected portfolio return. The set of tradeable transactions is executed.

What is claimed is: 
     
         1 . A machine learning portfolio generating 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, a portfolio construction request datastructure, the portfolio construction request datastructure structured to include a set of optimization parameters including a universe of securities, a time period length, a conditional value at risk portion, a conditional value at risk threshold, a portfolio value amount; 
 determine, via at least one processor, a set of simulated market scenarios associated with the time period length, the set of simulated market scenarios generated using a set of deep learning neural networks, each simulated market scenario in the set of simulated market scenarios structured to comprise a set of simulated market factor values; 
 retrieve, via at least one processor, a set of expected returns for securities in the universe of securities for the set of simulated market scenarios, each expected return in the set of expected returns configured as calculated for a security during a simulated market scenario using:
 the respective security'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 security, based on a first subset of the set of simulated market factor values, and 
 the respective security'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 security, based on a second subset of the set of simulated market factor values; 
 
 optimize, via at least one processor, portfolio weights of securities in the universe of securities in accordance with the conditional value at risk portion, the conditional value at risk threshold, and the portfolio value amount, using the set of expected returns, to generate a set of tradeable transactions that maximize expected portfolio return of an optimized portfolio; and 
 execute, via at least one processor, the set of tradeable transactions to generate the optimized portfolio. 
   
     
     
         2 . The apparatus of  claim 1 , further, comprising:
 the instructions to generate the set of simulated market scenarios using the set of deep learning neural networks are structured to comprise instructions to:
 determine, via at least one processor, a set of historical market scenarios and a set of time period buckets; 
 determine, via at least one processor, for each time period bucket from the set of time period buckets, a subset of historical market scenarios, from the set of historical market scenarios, associated with the respective time period bucket; 
 train, via at least one processor, for each time period bucket from the set of time period buckets, a deep learning neural network, from the set of deep learning neural networks, using the subset of historical market scenarios associated with the respective time period bucket; 
 determine, via at least one processor, for each time period bucket from the set of time period buckets, a number of simulated market scenarios to generate using the trained deep learning neural network associated with the respective time period bucket; and 
 generate, via at least one processor, for each time period bucket from the set of time period buckets, the determined number of simulated market scenarios for the respective time period bucket, using the trained deep learning neural network associated with the respective time period bucket. 
   
     
     
         3 . The apparatus of  claim 2 , further, comprising:
 the instructions to determine the set of historical market scenarios are structured to comprise instructions to:
 determine, via at least one processor, a historical data set, a rolling window period length, and a set of market factors; 
 determine, via at least one processor, a set of rolling window periods using the historical data set and the rolling window period length; and 
 calculate, via at least one processor, for each market factor from the set of market factors, for each rolling window period from the set of rolling window periods, a change to the respective market factor during the respective rolling window period,
 each historical market scenario from the set of historical market scenarios structured to comprise calculated changes to the set of market factors during a rolling window period. 
 
   
     
     
         4 . The apparatus of  claim 3 , further, comprising:
 the instructions to calculate a change to a market factor during a rolling window period are structured to comprise instructions to:
 determine, via at least one processor, the delta between values of the market factor at a beginning time point and an ending time point of the rolling window period. 
   
     
     
         5 . The apparatus of  claim 4 , further, comprising:
 the processor-executable instructions structured as:
 determine, via at least one processor, that historical data for the market factor during the rolling window period is unavailable for a time point; and 
 impute, via at least one processor, the unavailable historical data for the time point using a machine learning method. 
   
     
     
         6 . The apparatus of  claim 3 , further, comprising:
 the rolling window period length is structured to be equal to the time period length.   
     
     
         7 . The apparatus of  claim 2 , further, comprising:
 the set of time period buckets is structured to have an equal fixed length for each time period bucket.   
     
     
         8 . The apparatus of  claim 2 , further, comprising:
 the set of time period buckets is structured to have a variable length for each time period bucket, the variable length for each time period bucket determined by judging the overall goodness of fit between the set of simulated market scenarios and the set of historical market scenarios.   
     
     
         9 . The apparatus of  claim 2 , further, comprising:
 the instructions to train a deep learning neural network for a time period bucket using the associated subset of historical market scenarios are structured to comprise instructions to:
 select, via at least one processor, a historical market scenario from the associated subset of historical market scenarios; and 
 train, via at least one processor, the deep learning neural network for the time period bucket on the selected historical market scenario using a variational autoencoder. 
   
     
     
         10 . The apparatus of  claim 9 , further, comprising:
 the deep learning neural network for the time period bucket is trained to generate a set of Gaussian mixture latent variables.   
     
     
         11 . The apparatus of  claim 10 , further, comprising:
 the instructions to generate simulated market scenarios for the time period bucket, using the trained deep learning neural network associated with the time period bucket, are structured to comprise instructions to:
 generate, via at least one processor, a set of random values for the set of Gaussian mixture latent variables; and 
 generate a simulated market scenario, from the simulated market scenarios for the time period bucket, from the generated set of random values using a neural network decoder of the trained deep learning neural network associated with the time period bucket. 
   
     
     
         12 . The apparatus of  claim 1 , further, comprising:
 the processor-executable instructions structured as:
 filter, via at least one processor, the set of simulated market scenarios associated with the time period length based on specified ranges of allowable values for specified customized market factors. 
   
     
     
         13 . The apparatus of  claim 1 , further, comprising:
 the processor-executable instructions structured as:
 filter, via at least one processor, the set of simulated market scenarios associated with the time period length based on specified business cycle settings. 
   
     
     
         14 . The apparatus of  claim 1 , further, comprising:
 the processor-executable instructions structured as:
 initialize, via at least one processor, starting portfolio weights of securities in the universe of securities to benchmark portfolio weights of a benchmark portfolio. 
   
     
     
         15 . The apparatus of  claim 1 , further, comprising:
 the portfolio weights of securities in the universe of securities are structured to be optimized by finding a mixed integer linear programming portfolio solution.   
     
     
         16 . The apparatus of  claim 1 , further, comprising:
 the set of simulated market scenarios is further generated using a set of multi-variate mixture datastructures.   
     
     
         17 . A machine learning portfolio generating processor-readable, non-transient medium, comprising processor-executable instructions structured as:
 obtain, via at least one processor, a portfolio construction request datastructure, the portfolio construction request datastructure structured to include a set of optimization parameters including a universe of securities, a time period length, a conditional value at risk portion, a conditional value at risk threshold, a portfolio value amount;   determine, via at least one processor, a set of simulated market scenarios associated with the time period length, the set of simulated market scenarios generated using a set of deep learning neural networks, each simulated market scenario in the set of simulated market scenarios structured to comprise a set of simulated market factor values;   retrieve, via at least one processor, a set of expected returns for securities in the universe of securities for the set of simulated market scenarios, each expected return in the set of expected returns configured as calculated for a security during a simulated market scenario using:
 the respective security'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 security, based on a first subset of the set of simulated market factor values, and 
 the respective security'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 security, based on a second subset of the set of simulated market factor values; 
   optimize, via at least one processor, portfolio weights of securities in the universe of securities in accordance with the conditional value at risk portion, the conditional value at risk threshold, and the portfolio value amount, using the set of expected returns, to generate a set of tradeable transactions that maximize expected portfolio return of an optimized portfolio; and   execute, via at least one processor, the set of tradeable transactions to generate the optimized portfolio.   
     
     
         18 . A machine learning portfolio generating 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, a portfolio construction request datastructure, the portfolio construction request datastructure structured to include a set of optimization parameters including a universe of securities, a time period length, a conditional value at risk portion, a conditional value at risk threshold, a portfolio value amount; 
 determine, via at least one processor, a set of simulated market scenarios associated with the time period length, the set of simulated market scenarios generated using a set of deep learning neural networks, each simulated market scenario in the set of simulated market scenarios structured to comprise a set of simulated market factor values; 
 retrieve, via at least one processor, a set of expected returns for securities in the universe of securities for the set of simulated market scenarios, each expected return in the set of expected returns configured as calculated for a security during a simulated market scenario using:
 the respective security'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 security, based on a first subset of the set of simulated market factor values, and 
 the respective security'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 security, based on a second subset of the set of simulated market factor values; 
 
 optimize, via at least one processor, portfolio weights of securities in the universe of securities in accordance with the conditional value at risk portion, the conditional value at risk threshold, and the portfolio value amount, using the set of expected returns, to generate a set of tradeable transactions that maximize expected portfolio return of an optimized portfolio; and 
 execute, via at least one processor, the set of tradeable transactions to generate the optimized portfolio. 
   
     
     
         19 . A machine learning portfolio generating processor-implemented process, comprising executing processor-executable instructions to:
 obtain, via at least one processor, a portfolio construction request datastructure, the portfolio construction request datastructure structured to include a set of optimization parameters including a universe of securities, a time period length, a conditional value at risk portion, a conditional value at risk threshold, a portfolio value amount;   determine, via at least one processor, a set of simulated market scenarios associated with the time period length, the set of simulated market scenarios generated using a set of deep learning neural networks, each simulated market scenario in the set of simulated market scenarios structured to comprise a set of simulated market factor values;   retrieve, via at least one processor, a set of expected returns for securities in the universe of securities for the set of simulated market scenarios, each expected return in the set of expected returns configured as calculated for a security during a simulated market scenario using:
 the respective security'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 security, based on a first subset of the set of simulated market factor values, and 
 the respective security'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 security, based on a second subset of the set of simulated market factor values; 
   optimize, via at least one processor, portfolio weights of securities in the universe of securities in accordance with the conditional value at risk portion, the conditional value at risk threshold, and the portfolio value amount, using the set of expected returns, to generate a set of tradeable transactions that maximize expected portfolio return of an optimized portfolio; and   execute, via at least one processor, the set of tradeable transactions to generate the optimized portfolio.