Back to results
US 10,257,116 B1Granted
Machine learning resource allocator
AssigneeTriangle IP, Inc.
InventorsSameer Vadera, Katherine S. Gaudry, Thomas D. Franklin
Life of this patent
Estimating the term…
Abstract
A method and system for allocation of resources is disclosed. A data source is mined to determine event vectors from a large number of cases that follow a branched processing model. Current event vectors are compared to the mined event vectors with machine learning to predict future nodes for the current event vectors. Historical resource allocations for the mined event vectors are used to determine resource allocation for the current event vector over time. Current event vectors are combined to produce a resource allocation curve showing past and future resources allocated.
Claims20
A resource management system for allocation of resources for event vectors using machine learning, the resource management system comprising:
- a data ingest server comprising a processor and memory with instructions configured to download data from a source, wherein the data is publically available,
- a vector processing server comprising a processor and memory with instructions configured to generate a dataset comprising a plurality of event vectors, wherein:
- the dataset is a function of the data,
- each of the plurality of event vectors comprise a plurality of nodes spaced in time according to events,
- each of the plurality of nodes corresponds to a branched processing model of possible nodes, wherein the branched processing model is characterized by possible event vectors mined from the data, and
- each of the plurality of event vectors traverses the branched processing model of possible nodes in a plurality different ways with different timing;
- a prediction processing server comprising a processor and memory with instructions configured to:
- load a first event vector corresponding to past events associated with a domain, wherein the plurality of event vectors are outside of the domain,
- machine match the first event vector to a first subset of the plurality event vectors,
- predict first future nodes for the first event vector through completion as a function of the machine matching, wherein a first predicted event vector is a function of the first event vector and the first future nodes,
- load a second event vector corresponding to past events associated with the domain,
- second machine match the second event vector to a second subset of the plurality event vectors,
- predict second future nodes for the second event vector through completion as a function of the machine matching, wherein a second predicted event vector is a function of the second event vector and the second future nodes;
- and
- an allocation server comprising a processor and memory with instructions configured to generate a resource allocation curve as a function of the first predicted event vector and the second predicted event vector to show resources allocated forward and backward in time.
12 more elements
A method for allocation of resources for event vectors using machine learning, the method comprising:
- downloading data from a source, wherein the data is publically available;
- generating a dataset comprising a plurality of event vectors, wherein:
- the dataset is a function of data,
- each of the plurality of event vectors comprise a plurality of nodes spaced in time according to events,
- each of the plurality of nodes is selected from a branched processing model of possible nodes, wherein the branched processing model is characterized by possible event vectors mined from the data, and
- each of the plurality of event vectors traverses the branched processing model of possible nodes in a plurality different ways with different timing;
- loading a first event vector corresponding to past events associated with a domain, wherein the plurality of event vectors are outside of the domain;
- loading a second event vector corresponding to past events associated with the domain;
- machine matching the first event vector to a subset of the plurality event vectors;
- machine matching the second event vector to a subset of the plurality event vectors;
- predicting first future nodes for the first event vector as a function of the machine matching, wherein a first predicted event vector is a function of the first event vector and the first future nodes;
- predicting second future nodes for the second event vector as a function of the machine matching, wherein a second predicted event vector is a function of the second event vector and the second future nodes;
- determining resource allocation curve as a function of the first predicted event vector and second predicted event vector to show resources allocated in the past and future.
10 more elements
A resource management system for allocation of resources for event vectors using machine learning, the resource management system comprising:
- one or more processors, and
- one or memories coupled with said one or more processors, wherein the one or more processors and one or more memories are configured to:
- download data from a source, wherein the data is publically available;
- generate a dataset comprising a plurality of event vectors, wherein:
- the dataset is a function of data,
- each of the plurality of event vectors comprise a plurality of nodes spaced in time according to events,
- each of the plurality of nodes is selected from a branched processing model of possible nodes, wherein the branched processing model is characterized by possible event vectors mined from the data, and
- each of the plurality of event vectors traverses the branched processing model of possible nodes in a plurality different ways with different timing;
- load a first event vector corresponding to past events associated with a domain, wherein the plurality of event vectors are outside of the domain;
- load a second event vector corresponding to past events associated with the domain;
- machine match the first event vector to a subset of the plurality event vectors;
- machine match the second event vector to a subset of the plurality event vectors;
- predict first future nodes for the first event vector as a function of the machine matching, wherein a first predicted event vector is a function of the first event vector and the first future nodes;
- predict second future nodes for the second event vector as a function of the machine matching, wherein a second predicted event vector is a function of the second event vector and the second future nodes;
- determine resource allocation curve as a function of the first predicted event vector and second predicted event vector to show resources allocated in the past and future.
12 more elements
Dependencies are inferred from the claim text.