US 12,141,822 B1
System and method to extrapolate inventory sales data using incomplete market data
Shmuel Sherman, Valley Stream, NY (US); and Yiren Li, New York, NY (US)
Assigned to BROKER GENIUS INC., Far Rockaway, NY (US)
Filed by BROKER GENIUS, INC., Far Rockaway, NY (US)
Filed on Jul. 27, 2020, as Appl. No. 16/939,806.
Claims priority of provisional application 62/878,930, filed on Jul. 26, 2019.
Int. Cl. G06Q 30/0201 (2023.01); G06Q 10/087 (2023.01)
CPC G06Q 30/0201 (2013.01) [G06Q 10/087 (2013.01)] 11 Claims
OG exemplary drawing
 
1. A real-time database filtering system comprising:
at least one database configured to store open transaction data;
at least one processor configured by code executing therein to:
access open transaction data from the at least one database for a first time period to generate a first transaction dataset;
generate a unique identifier for each transaction in the first transaction dataset wherein the unique identifier is based on the collection of data values associated with the transaction;
generate, using a predictive model, a resource access value, wherein the predictive model is a neural network trained on a training dataset of completed transaction data, where each element of the training dataset includes at least a location and a price transaction characteristic for a given transaction in the training dataset and a time duration value corresponding to the amount of time elapsed to complete the given transaction, wherein the predictive model is configured receive at least a portion of the open transaction data and output the resource access value;
access open transaction data from the database at a second time period to generate a second transaction dataset, wherein the second time period is determined using the generated resource access value such that accessing the open transaction at the second time period occurs only in intervals of the resource access value;
generate a unique identifier for each transaction in the second transaction dataset wherein the unique identifier is based on the same data associated with the transaction as used to create the unique identifiers for each transaction in the first transaction dataset;
determine a subset of the first transaction dataset by identifying each open transaction that is present in the first transaction dataset and not present in the second transaction dataset by comparing the unique identifiers present in the first transaction dataset and the unique identifiers present in the second transaction dataset;
assign a comparison score to each member of the subset of the first transaction dataset, wherein the comparison score for each entry in the subset of the first transaction dataset is determined, at least by, obtaining the difference between the price of the access credential for the entry in a predetermined area and the lowest price for an access credential in the first transaction database for the same predetermined area wherein the greater the difference obtained the lower the comparison score;
compare the comparison score for each member of the subset of the first transaction dataset to a threshold value,
classify, where the comparison score exceeds the threshold value, the member of the transaction data subset as a closed transaction; and
update a transaction database to include each member of the transaction data subset classified as a closed transaction.