Systems and methods for allocating fractional shares of a public offering

Abstract

A share allocation (SA) computing device includes a processor in communication with a database. The processor is configured to execute a computational model including a plurality of model layers. The plurality of model layers includes a fractional node layer configured to assign each candidate investor of a plurality of candidate investors to a corresponding node. Each node is associated with a weight, and the nodes define an interconnected neural network. The fractional node layer is also configured to apply a machine learning algorithm configured to adjust the weights of the nodes in response to respective fitness values input to the nodes, and convert the adjusted weight for each node into a corresponding fraction. The fractional node layer is further configured to allocate, to each candidate investor, a respective fractional share of an offering, the fractional share corresponding to the fraction associated with the corresponding node.

What is claimed is: 
     
         1 . A share allocation (SA) computing device comprising at least one processor in communication with a database, the at least one processor configured to:
 retrieve investor data from the database, wherein the investor data is associated with past investment activity of a plurality of individual investors;   for each investor of the plurality of individual investors, compute an investor score using the investor data;   assign each investor of the plurality of individual investors to a corresponding node of a plurality of nodes in a fractional node layer;   associate each corresponding node with a weight based on the investor score for the individual investor assigned to the corresponding node;   iteratively apply a heuristic algorithm to subsets of the plurality of nodes to adjust upward the weight of at least one node in at least one subset of the subsets based at least in part on the investor score associated with the at least one node as compared to investor scores associated with other nodes in the at least one subset;   output updated weights for the plurality of nodes based on an output from the heuristic algorithm; and   allocate, to each investor of the plurality of individual investors, a respective fractional share of an offering, wherein the respective fractional shares correspond to the updated weights associated with the plurality of nodes.   
     
     
         2 . The SA computing device of  claim 1 , wherein the at least one processor is further configured to:
 receive an input identifying a first allocation of shares of the offering among the plurality of individual investors, the first allocation of shares including a respective number of whole shares allocated to each of the individual investors; and   output a second allocation of shares of the offering among the plurality of individual investors, the second allocation of shares including, for each of the plurality of individual investors, the respective number of whole shares from the first allocation of shares adjusted by the respective fractional shares.   
     
     
         3 . The SA computing device of  claim 1 , wherein the at least one processor is further configured to:
 access request data for each individual investor indicating a stated amount the individual investor is willing to spend on the offering; and   constrain the weight for each node such that a purchase price of the shares in the allocation to the corresponding individual investor does not exceed the stated amount.   
     
     
         4 . The SA computing device of  claim 1 , wherein the at least one processor is further configured to receive at least a portion of the investor data from at least one broker-dealer computing device, wherein the received at least portion of the investor data is associated with past investment transactions of the plurality of individual investors conducted through channels external to the SA computing device. 
     
     
         5 . The SA computing device of  claim 1 , wherein the at least one processor is further configured to:
 capture data from investment transactions of the plurality of individual investors conducted through the SA computing device; and   store the captured data in the database as at least a portion of the investor data.   
     
     
         6 . The SA computing device of  claim 1 , wherein the investor data includes (i) external data fields associated with past investment transactions of the plurality of individual investors conducted through channels external to the SA computing device, and (ii) internal data fields associated with past investment transactions of the plurality of individual investors conducted through the SA computing device, and wherein the at least one processor is further configured to:
 calculate an external vector using external factors derived from the external data fields;   calculate an internal vector using internal factors derived from the internal data fields; and   compute the investor score for each individual investor using a weighted combination of the external vector and the internal vector.   
     
     
         7 . The SA computing device of  claim 6 , wherein the at least one processor is further configured to calculate the external vector as a weighted average of the external factors. 
     
     
         8 . The SA computing device of  claim 6 , wherein the at least one processor is further configured to calculate the internal vector as a weighted average of the internal factors. 
     
     
         9 . The SA computing device of  claim 6 , wherein the offering is associated with an industry classification, and wherein the at least one processor is further configured to:
 query the database for investor data records in which the past investment activity related to the industry classification; and   use only the investor data records returned from the database to calculate at least one of the external vector or the internal vector.   
     
     
         10 . A computer-implemented method, the method implemented by a share allocation (SA) computing device comprising at least one processor in communication with a database, the method comprising:
 retrieving investor data from the database, wherein the investor data is associated with past investment activity of a plurality of individual investors;   for each investor of the plurality of individual investors, computing an investor score using the investor data;   assigning each investor of the plurality of individual investors to a corresponding node of a plurality of nodes in a fractional node layer;   associating each corresponding node with a weight based on the investor score for the individual investor assigned to the corresponding node;   iteratively applying a heuristic algorithm to subsets of the plurality of nodes to adjust upward the weight of at least one node in at least one subset of the subsets based at least in part on the investor score associated with the at least one node as compared to investor scores associated with other nodes in the at least one subset;   outputting updated weights for the plurality of nodes based on an output from the heuristic algorithm; and   allocating, to each investor of the plurality of individual investors, a respective fractional share of an offering, wherein the respective fractional shares correspond to the updated weights associated with the plurality of nodes.   
     
     
         11 . The computer-implemented method of  claim 10 , further comprising:
 receiving an input identifying a first allocation of shares of the offering among the plurality of individual investors, the first allocation of shares including a respective number of whole shares allocated to each of the individual investors; and   outputting a second allocation of shares of the offering among the plurality of individual investors, the second allocation of shares including, for each of the plurality of individual investors, the respective number of whole shares from the first allocation of shares adjusted by the respective fractional shares.   
     
     
         12 . The computer-implemented method of  claim 10 , further comprising:
 accessing request data for each individual investor indicating a stated amount the individual investor is willing to spend on the offering; and   constraining the weight for each node such that a purchase price of shares in the allocation to the corresponding individual investor does not exceed the stated amount.   
     
     
         13 . The computer-implemented method of  claim 10 , wherein the offering is associated with an industry classification, the method further comprising:
 querying the database for investor data records in which the past investment activity related to the industry classification; and   using only the investor data records returned from the database to compute the investor scores.   
     
     
         14 . At least one non-transitory computer-readable storage media having computer-executable instructions embodied thereon, wherein when executed by share allocation (SA) computing device having at least one processor in communication with a database, the computer-executable instructions cause the SA computing device to:
 retrieve investor data from the database, wherein the investor data is associated with past investment activity of a plurality of individual investors;   for each investor of the plurality of individual investors, compute an investor score using the investor data;   assign each investor of the plurality of individual investors to a corresponding node of a plurality of nodes in a fractional node layer;   associate each corresponding node with a weight based on the investor score for the individual investor assigned to the corresponding node;   iteratively apply a heuristic algorithm to subsets of the plurality of nodes to adjust upward the weight of at least one node in at least one subset of the subsets based at least in part on the investor score associated with the at least one node as compared to investor scores associated with other nodes in the at least one subset;   output updated weights for the plurality of nodes based on an output from the heuristic algorithm; and   allocate, to each investor of the plurality of individual investors, a respective fractional share of an offering, wherein the respective fractional shares correspond to the updated weights associated with the plurality of nodes.   
     
     
         15 . The at least one non-transitory computer-readable storage media of  claim 14 , wherein the computer-executable instructions further cause the SA computing device to:
 receive an input identifying a first allocation of shares of the offering among the plurality of individual investors, the first allocation of shares including a respective number of whole shares allocated to each of the individual investors; and   output a second allocation of shares of the offering among the plurality of individual investors, the second allocation of shares including, for each of the plurality of individual investors, the respective number of whole shares from the first allocation of shares adjusted by the respective fractional shares.   
     
     
         16 . The at least one non-transitory computer-readable storage media of  claim 14 , wherein the computer-executable instructions further cause the SA computing device to:
 access request data for each individual investor indicating a stated amount the individual investor is willing to spend on the offering; and   constrain the weight for each node such that a purchase price of shares in the allocation to the corresponding individual investor does not exceed the stated amount.   
     
     
         17 . The at least one non-transitory computer-readable storage media of  claim 14 , wherein the computer-executable instructions further cause the SA computing device to receive at least a portion of the investor data from at least one broker-dealer computing device, wherein the received at least portion of the investor data is associated with past investment transactions of the plurality of individual investors conducted through channels external to the SA computing device. 
     
     
         18 . The at least one non-transitory computer-readable storage media of  claim 14 , wherein the computer-executable instructions further cause the SA computing device to:
 capture data from investment transactions of the plurality of individual investors conducted through the SA computing device; and   store the captured data in the database as at least a portion of the investor data.   
     
     
         19 . The at least one non-transitory computer-readable storage media of  claim 14 , wherein the plurality of individual investors is a subset of a second plurality of individual investors, and the at least one processor is further configured to:
 receive an input identifying the second plurality of individual investors; and   select, for assignment to the corresponding nodes, the subset of individual investors from among the second plurality of individual investors.   
     
     
         20 . The at least one non-transitory computer-readable storage media of  claim 14 , wherein the offering is associated with an industry classification, and wherein the computer-executable instructions further cause the SA computing device to:
 query the database for investor data records in which the past investment activity related to the industry classification; and   use only the investor data records returned from the database to compute the investor scores.