Creation, use and training of computer-based discovery avatars

Abstract

In embodiments of the present invention improved capabilities are described for developing, training, validating and deploying discovery avatars embodying mathematical models that may be used for document and data discovery and deployed within large data repositories.

1 . A method of constructing a computer-based discovery avatar, embodied in a non-transitory computer readable medium that manages a queue of data stream elements to aid an investigation and, when executing on one or more computers, performs the following steps:
 Step One; tokenizing source data;   extracting a plurality of data features from the tokenized source data, wherein the extracted data features are stored as quantitative vectors;   Step Two: analyzing the extracted data features using a mathematical model to determine a data cluster, wherein the data cluster includes extracted data features that share an attribute and includes identifiers that are associated with a plurality of data elements from the source data;   presenting a first source datum for review, from the plurality of data elements from the source data, based at least in part on the identifiers within the data cluster;   scoring the first source datum based at least in part on its relevance to a substantive topic;   presenting a second source datum for review, from the plurality of data elements from the source data, based at least in part on the identifiers within the data cluster;   scoring the second source datum based at least in part on its relevance to the substantive topic;   comparing the score of the first source datum to the score of the second source datum;   optimizing the mathematical model based at least in part on the comparison of scores;   Step Three: iteratively performing Step Two to improve scores received by data elements from the source data that are selected using the mathematical model to create an optimized model; and   storing the optimized model as a computer-based discovery avatar.   
     
     
         2 . The method of  claim 1 , wherein the source data is a stored repository of documents. 
     
     
         3 . The method of  claim 1 , wherein the source data derives from a plurality of distributed data storage repositories. 
     
     
         4 . The method of  claim 1 , wherein the tokenization is white space tokenization. 
     
     
         5 . The method of  claim 1 , wherein the scoring is performed by a human. 
     
     
         6 . The method of  claim 5 , wherein the scoring by the human is quantitatively weighted by a metadatum associated with the human. 
     
     
         7 . The method of  claim 6 , wherein the metadatum is a job title. 
     
     
         8 . The method of  claim 6 , wherein the metadatum is a credential. 
     
     
         9 . The method of  claim 1 , wherein the scoring is performed by at algorithm. 
     
     
         10 . The method of  claim 1 , wherein the discovery avatar categorizes the source data based at least in part on the use of support vector machines. 
     
     
         11 . The method of  claim 1 , wherein the discovery avatar is deployed for use on a second data source to create a second set of data clusters using the optimized model of the discovery avatar. 
     
     
         12 . The method of  claim 1 , wherein the discovery avatar is deployed for use on a plurality of data sources to create a plurality of data clusters that are scored and used to rank each of the plurality of data sources according to relevance to the substantive topic. 
     
     
         13 . A method of constructing a family of computer-based discovery avatars, each embodied in a non-transitory computer readable medium that manages a queue of data stream elements to aid an investigation and, when executing on one or more computers, performs the following steps:
 Step One: tokenizing source data;   extracting a plurality of data features from the tokenized source data, wherein the extracted data features are stored as quantitative vectors;   Step Two: analyzing the extracted data features using a mathematical model to determine a data cluster, wherein the data cluster includes extracted data features that share an attribute that is related to a super-set topic, and includes identifiers that are associated with a plurality of data elements from the source data;   presenting and scoring data elements from the source data based at least in part on the identifiers within the data cluster relating to the super-set topic;   optimizing the mathematical model based at least in part on a comparison of the scored data elements;   Step Three: storing the optimized model as a computer-based discovery avatar parent;   Step Four: repeating Step One and Step Two, wherein the data cluster includes a second set of extracted data features that share a second attribute that is related to both the super-set topic and a subset topic, and results in a second optimized model that is based on the super-set and subset topics and is stored as a computer-based discovery avatar child.   
     
     
         14 . The method of  claim 13 , wherein the subset topic is defined by terms that are included in a set of terms used to define the super-set topic. 
     
     
         15 . The method of  claim 13 , wherein the subset topic is defined by terms that are additive to a set of terms used to define the super-set topic. 
     
     
         16 . The method of  claim 13 , wherein the avatar parent is memorialized and locked from further iterative improvement. 
     
     
         17 . Wherein the parent (alternatively, child) avatar is deployed as an analytic commodity for use on a third source of data. 
     
     
         18 . Wherein the genealogy of avatar parent-avatar child relations is presented in a graphic user interface. 
     
     
         19 . A method of training a computer-based discovery avatar using a second computer-based discovery avatar, embodied in a non-transitory computer readable medium that, when executing on one or more computers, performs the following steps:
 identifying at least one attribute of a first mathematical model inherent in a first computer-based discovery avatar that is relevant to a second mathematical model inherent in a second computer-based discovery avatar;   incorporating a second attribute from the first mathematical model inherent in the first computer-based discovery avatar within the second computer-based discovery avatar to create a cross-trained mathematical model in the second computer-based discovery avatar; and   validating the cross-trained mathematical model by deploying the second computer-based discovery avatar on a set of source data substantially similar to source data on which the first computer-based avatar was developed, wherein the validation is confirmed based at least in part on a comparison of data clusters derived using the first discovery avatar and data clusters derived using the cross-trained mathematical model of the second computer-based discovery avatar.   
     
     
         20 . The method of  claim 19 , wherein the relevance of the at least one attribute is based at least in part on a quantitative association to a substantive topic inherent to a data source.