Systems and techniques for predictive data analytics

Patent metadata

Field	Value
Publication number	US-9489630-B2
Application number	US-201514720079-A
Country	US
Kind code	B2
Filing date	May 22, 2015
Priority date	May 23, 2014
Publication date	Nov 8, 2016
Grant date	Nov 8, 2016

Abstract

Systems and techniques for predictive data analytics are described. In a method for selecting a predictive model for a prediction problem, the suitabilities of predictive modeling procedures for the prediction problem may be determined based on characteristics of the prediction problem and/or on attributes of the respective modeling procedures. A subset of the predictive modeling procedures may be selected based on the determined suitabilities of the selected modeling procedures for the prediction problem. A resource allocation schedule allocating computational resources for execution of the selected modeling procedures may be generated, based on the determined suitabilities of the selected modeling procedures for the prediction problem. Results of the execution of the selected modeling procedures in accordance with the resource allocation schedule may be obtained. A predictive model for the prediction problem may be selected based on those results.

First claim

What is claimed is:

1. A method for selecting a predictive model for a prediction problem, the method comprising:
determining suitabilities of a plurality of predictive modeling procedures for the prediction problem based, at least in part, on characteristics of the prediction problem and/or on attributes of the respective modeling procedures;
selecting at least a subset of the predictive modeling procedures based, at least in part, on the determined suitabilities of the selected modeling procedures for the prediction problem;
transmitting instructions to a plurality of processing nodes, the instructions comprising a resource allocation schedule allocating resources of the processing nodes for execution of the selected modeling procedures, the resource allocation schedule being based, at least in part, on the determined suitabilities of the selected modeling procedures for the prediction problem;
receiving results of the execution of the selected modeling procedures by the plurality of processing nodes in accordance with the resource allocation schedule, wherein the results include predictive models generated by the selected modeling procedures, and/or scores of the models for data associated with the prediction problem; and
selecting, from the generated predictive models, a predictive model for the prediction problem based, at least in part, on the score of the predictive model,
wherein determining the suitabilities of the plurality of predictive modeling procedures for the prediction problem comprises determining the suitability of a first of the plurality of predictive modeling procedures for the prediction problem, including:
selecting one or more prediction problems based, at least in part, on similarity between characteristics of the prediction problem and characteristics of the one or more prediction problems;
selecting one or more predictive modeling procedures based, at least in part, on similarity between the first predictive modeling procedure and the one or more predictive modeling procedures; and
processing data indicative of results of applying the one or more predictive modeling procedures to the one or more prediction problems.

2. The method of claim 1 , wherein determining the suitabilities of the plurality of predictive modeling procedures for the prediction problem comprises eliminating at least one predictive modeling procedure from consideration based on one or more relationships between the characteristics of the prediction problem and the attributes of the eliminated procedure.

3. The method of claim 1 , wherein determining the suitabilities of the plurality of predictive modeling procedures for the prediction problem comprises assigning a suitability value to at least one predictive modeling procedure based on one or more relationships between the characteristics of the prediction problem and the attributes of the at least one predictive modeling procedure.

4. The method of claim 1 , wherein selecting the one or more prediction problems based, at least in part, on the similarity between characteristics of the prediction problem and characteristics of the one or more prediction problems comprises selecting the one or more prediction problems based, at least in part, on similarity between characteristics of the data associated with the prediction problem and characteristics of data associated with the one or more prediction problems.

5. The method of claim 1 , wherein determining the suitability of the first predictive modeling procedure for the prediction problem further comprises determining the similarity between the first modeling procedure and the one or more modeling procedures based, at least in part, on processing steps performed by the first modeling procedure and the one or more modeling procedures.

6. The method of claim 1 , wherein processing the data indicative of the results of applying the one or more modeling procedures to the one or more prediction problems comprises predicting the suitability of the first predictive modeling procedure for the prediction problem by applying a second predictive modeling procedure to the data indicative of the results of applying the one or more modeling procedures to one or more prediction problems.

7. The method of claim 6 , further comprising:
using the selected predictive model to predict outcomes of instances of the prediction problem, wherein the selected predictive model is generated by a particular predictive modeling procedure included in the selected subset of the predictive modeling procedures; and
updating data indicative of results of applying the particular predictive modeling procedure to the prediction problem based, at least in part, on a relationship between the predicted outcomes and actual outcomes of the instances of the prediction problem.

8. The method of claim 1 , wherein determining the suitabilities of the plurality of predictive modeling procedures comprises assigning suitability scores to the respective modeling procedures included in the plurality of predictive modeling procedures, and wherein selecting at least a subset of the predictive modeling procedures comprises selecting, from the plurality of predictive modeling procedures, one or more predictive modeling procedures having suitability scores that exceed a threshold suitability score.

9. The method of claim 8 , further comprising determining the threshold suitability score based, at least in part, on an amount of processing resources available for execution of the selected subset of the modeling procedures.

10. The method of claim 1 , wherein determining the suitabilities of the plurality of predictive modeling procedures comprises assigning suitability scores to the respective modeling procedures included in the plurality of predictive modeling procedures, and wherein selecting at least a subset of the predictive modeling procedures comprises selecting, from the plurality of predictive modeling procedures, one or more predictive modeling procedures having suitability scores within a specified range of a highest suitability score assigned to any of the predictive modeling procedures for the prediction problem.

11. The method of claim 1 , wherein selecting at least a subset of the predictive modeling procedures comprises selecting, from the plurality of predictive modeling procedures, approximately a specified fraction of the predictive modeling procedures having highest suitability scores.

12. The method of claim 1 , wherein selecting at least a subset of the predictive modeling procedures comprises selecting at least one predictive modeling procedure based, at least in part, on user input.

13. The method of claim 1 , wherein the allocated resources of the processing nodes comprise execution cycles of the processing nodes, execution time on the processing nodes, and/or computer-readable storage of the processing nodes.

14. The method of claim 1 , wherein the processing nodes comprise one or more cloud-based processing nodes.

15. The method of claim 1 , wherein the selected subset of modeling procedures comprises first and second selected modeling procedures determined to have first and second suitabilities for the prediction problem, respectively, the first suitability of the first selected modeling procedure being greater than the second suitability of the second selected modeling procedure, and wherein the resource allocation schedule allocates resources of the processing nodes to the first and second selected modeling procedures based, at least in part, on the first and second suitabilities.

16. The method of claim 15 , wherein the resource al

Assignees

• Datarobot Inc

Inventors

• Achin Jeremy
• Degodoy Thomas
• Owen Timothy
• Conort Xavier

Classifications

• G06F16/28

  Databases characterised by their database models, e.g. relational or object models · CPC title

• G06Q10/06

  Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling · CPC title

• G06N5/02

  Knowledge representation; Symbolic representation · CPC title

• G06N5/04 Primary

  Inference or reasoning models · CPC title

• G06F9/5011

  the resources being hardware resources other than CPUs, Servers and Terminals · CPC title