Systems and methods for training machine learning models using generated faceted models

1 . A system for a training a machine learning model with a hybrid dataset, comprising:

a training server;

a matter database server;

a data source database server; and

a computer network, wherein:

said training server, said matter database server and said data source database server are in communication through said computer network;

said training server comprises a hybrid dataset engine and a machine learning model engine;

said hybrid dataset engine constructs said hybrid dataset from data stored in said matter database server and said data source database server;

said machine learning model engine trains said machine learning model according to said hybrid dataset, to form a faceted machine learning model;

said hybrid dataset is constructed according to exclusionary criteria, such that said faceted machine learning model analyzes data after training that is not in said hybrid dataset;

said matter database comprises a plurality of records to be analyzed by said faceted machine learning model;

said records are not present in said hybrid dataset; and

said faceted machine learning model analyzes said records to determine a content of said records.

2 . The system of claim 1 , wherein said hybrid dataset engine constructs said hybrid dataset by removing a number N of records from data source database server, selecting the same number N of records from the matter database server, and replacing said N records from data source database server with the N records from said matter database server to create said hybrid dataset.

3 . The system of claim 1 , wherein said hybrid dataset engine constructs said hybrid dataset by adding a number N of records from said matter database server to said records of said data source database server.

4 . The system of claim 3 , wherein said hybrid dataset engine selects data from said matter database server according to one or more matter filtration criteria.

5 . The system of claim 4 , wherein said matter filtration criteria comprises a filtration selected from the group consisting of time, channel, subject matter, and person.

6 . The system of claim 5 , wherein said channel comprises records within a particular communication medium, between communication media or a combination thereof.

7 . The system of claim 6 , wherein said communication media is selected from the group consisting of chat and messaging software, emails, letters sent on paper or other physical media, and recordings and/or transcripts of voice communication.

8 . The system of claim 7 , wherein said communication media permits channels, group messages and/or threads, and wherein said filtration is further based according to one or more of said channels, message groups and/or threads.

9 . The system of claim 5 , wherein said matter filtration criteria are exclusionary, such that data to be included in analysis of said matter by said faceted machine learning model are not included in said hybrid dataset.

10 . The system of claim 9 , wherein said server further comprises a matter ingestion engine and wherein said matter ingestion engine applies said one or more matter filtration criteria.

11 . The system of claim 10 , wherein;

said matter ingestion engine divides records from said matter database server according to said one or more matter filtration criteria,

said records are divided into a plurality of groups, including at least one group of records meeting said one or more matter filtration criteria, at least one group of records meeting a majority of said one or more matter filtration criteria, and at least one group of records meeting a minority of said one or more matter filtration criteria;

said group of majority records form a primary sideband and said group of minority records form a secondary sideband; and

said hybrid dataset engine selects at least some records from said primary sideband.

12 . The system of claim 11 , wherein said hybrid dataset engine selects at least some records from said secondary sideband.

13 . The system of claim 5 , wherein said hybrid dataset engine further randomly selects data from within data stored at said matter database server that is selected according to one or more matter filtration criteria.

14 . The system of claim 3 , wherein said hybrid dataset engine selects data from said matter database server at random.

15 . The system of claim 1 , wherein said hybrid dataset engine constructs said hybrid dataset by adding or removing a number N of records from said matter database server to said records of said data source database server, according to a size of dataset available at said matter database server, relative to an original training dataset size for said faceted machine learning model.

16 . The system of claim 1 , wherein said faceted machine learning model flags content that is illegal, unethical, against company policy or a combination thereof; and wherein said faceted machine learning model outputs a warning in regard to said flagged content.

17 . The system of claim 16 , further comprising a user computational device, operating a user app interface, in communication with said faceted machine learning model through said computer network; wherein said records are analyzed by said faceted machine learning model according to a request through said user app interface.

18 . The system of claim 17 , wherein said request is made for real time record analysis.

19 . The system of claim 17 , wherein said request is for ongoing monitoring of records generated by one or more communication media.

20 . The system of claim 17 , further comprising a server gateway in communication with said training server, said matter database server, and said data source database server; wherein said server gateway operates said faceted machine learning model.