Systems and methods for cloud content-based document clustering and classification integration

Inventors

Zeng, Zhihong • Chaudhry, Anwar • Chandrasekhar, Rajesh • Williams, Adam Darius • Gandhi, Utpal N.

Assignees

Iron Mountain Inc

Abstract

A computer-implemented method, includes accessing, by a processor, a set of asset documents. The method also includes performing, by the processor, feature extraction on text of each document of the set of asset documents using a feature extraction module to generate a set of features, where each feature of the set of features represents a document of the set of asset documents. The method also includes generating, by the processor, a set of lower-dimensional features from the set of features using a singular value decomposition module. The method also includes generating, by the processor, a set of clusters from the set of lower-dimensional features using a clustering module. The method also includes training, by the processor, a machine-learning model of a classification microservice using the set of clusters generated from the clustering module.

Core Innovation

A computer-implemented method accesses a set of asset documents and receives a feature extraction tool selection from a set of feature extraction tools stored within a feature extraction module. The selection comprises a bidirectional encoder representation of transformers (BERT) based feature extraction tool or a bag-of-words feature extraction tool. The processor performs feature extraction on text of each document of the set of asset documents using the selected tool to generate a set of features, where each feature represents a document of the set of asset documents.

The processor provides the set of features to a singular value decomposition module and generates a set of lower-dimensional features from the set of features using the singular value decomposition module. The processor then generates a set of clusters representing document classes from the set of lower-dimensional features using a clustering module. The clustering module is configured to perform a density based spatial clustering operation by assigning at least one document of the set of asset documents as a border sample document within the cluster of the set of clusters.

The processor performs a grid search to adjust adjustable parameters of the feature extraction module, the singular value decomposition module, the clustering module, or a combination thereof. The processor trains a machine-learning model of a classification microservice using the set of clusters generated from the clustering module. The classification microservice assigns a document class of the set of document classes to additional documents added to the set of asset documents and further trains the machine-learning model using results of clustering the additional documents.

Claims Coverage

The independent claims are clm-00001, clm-00005, and clm-00008. Each independent claim includes multiple inventive features, with core features spanning feature extraction tool selection, singular value decomposition to generate lower-dimensional features, density based spatial clustering with border sample assignment to form clusters representing document classes, and grid search followed by supervised classification training in a classification microservice.

Across independent claims, the inventive coverage centers on combining selectable text feature extraction, singular value decomposition to form lower-dimensional features, density based spatial clustering with border sample document assignment to produce clusters representing document classes, using grid search to adjust module parameters, and training a classification microservice that assigns classes to additional documents and further trains using clustering results.

Stated Advantages

Documented Applications