Comparing projects

Patent metadata

Field	Value
Publication number	US-9928239-B2
Application number	US-201414559053-A
Country	US
Kind code	B2
Filing date	Dec 3, 2014
Priority date	Apr 9, 2010
Publication date	Mar 27, 2018
Grant date	Mar 27, 2018

Abstract

A project-comparator subsystem generates a similarity metric for input electronically-represented projects input to the project-comparator subsystem. A project search engine receives data that represents a first electronically-represented project, identifies, by pairwise comparison of the stored electronically-represented projects with the first electronically-represented project using the project-comparator subsystem, a stored electronically-represented project with greatest similarity to the first electronically-represented project, and outputs an indication of the identified stored electronically-represented project.

First claim

The invention claimed is:

1. A system comprising:
at least one non-transitory computer-readable storage medium to store electronically-represented projects represented by computer-accessible data;
at least one processor;
a project-comparator subsystem executable on the at least one processor, wherein the project-comparator subsystem comprises at least one first level feature comparator and a second level comparator;
a project search engine executable on the at least one processor to:
receive data that represents a first electronically-represented project;
invoke the project-comparator subsystem to compare the received first electronically-represented project to at least one of the stored electronically-represented projects, wherein to compare the received first electronically-represented project to at least one of the stored electronically-represented projects, the project-comparator subsystem is to:
extract features from the first electronically-represented project and the at least one of the stored electronically-represented projects;
determine, by the at least one first-level comparator, a similarity metric for each of the extracted features; and
determine, by the second-level comparator, an overall similarity metric, by applying weights to the similarity metrics for the extracted features, and computationally combing the weighted similarity metrics; and

a learning module executable on the at least one processor to:
receive a feedback regarding similarity between the first electronically-represented project and the at least one of the stored electronically-represented projects; and

compute adjustments to the second-level comparator in response to the received feedback.

2. The system of claim 1 , wherein the received feedback is from a user or a program.

3. The system of claim 1 , wherein the received feedback relates to an accuracy of identification of the electronically-represented projects as being similar by the project search engine.

4. The system of claim 1 , wherein the adjustments comprise adjustments of tunable parameters used by the project-comparator subsystem to generate the similarity metrics.

5. The system of claim 4 , wherein the learning module is to:
identify a first subset of the features associated with a positive feedback, and identify a second subset of the features associated with a negative feedback; and
modify the tunable parameters based on the first and second subsets of the features.

6. The system of claim 5 , wherein the modifying of the tunable parameters is to ameliorate correlation with the second subset of the features, and to produce correlation with the first subset of the features.

7. The system of claim 1 , wherein the at least one first level feature comparator includes feature comparators to compare an extracted text from the stored electronically-represented projects by creating a term-occurrence vector from the extracted text for each stored electronically-represented project and computing a cosine of an angle between the term-occurrence vectors.

8. The system of claim 1 , wherein the at least one first level feature comparator includes feature comparators to compare numerical values extracted from the stored electronically-represented projects.

9. The system of claim 1 , wherein the at least one first level feature comparator includes feature comparators to compare a figure or graphical data extracted from the stored electronically-represented projects by rotating and scaling the figure or graphical data extracted from a first of the stored electronically-represented projects with respect to the figure or graphical data extracted from a second of the stored electronically-represented projects, and to compute a metric corresponding to an overlap of features within the figure or graphical data extracted from the first of the stored electronically-represented projects and the figure or graphical data extracted from the second of the stored electronically-represented projects.

10. The system of claim 1 wherein the learning module is to be repeatedly invoked at respective points in time.

11. A method comprising:
receiving an electronically-represented project by a project search engine implemented on one or more computer systems;
comparing, by a project-comparator subsystem, the received electronically-represented project to at least one of stored electronically-represented projects, wherein the comparing comprises:
extracting features from the received electronically-represented project and the at least one of the stored electronically-represented projects;
determining, by at least one first-level comparator of the project-comparator subsystem, a similarity metric for each of the extracted features; and
determining, by a second-level comparator project-comparator subsystem, an overall similarity metric, by applying weights to the similarity metric for the extracted features, and computationally combining the weighted similarity metrics; and

receiving, by a learning module, a feedback regarding similarity between the received electronically-represented project and the at least one of the stored electronically-represented projects; and
computing, by the learning module, adjustments to the second-level comparator in response to the received feedback.

12. The method of claim 11 , wherein the adjustments comprise adjustments of tunable parameters used by the project-comparator subsystem to generate the similarity metric for the at least one of the stored electronically-represented projects.

13. The system of claim 12 , further comprising:
identifying, by the learning module, a first subset of the features associated with a positive feedback;
identifying, by the learning module, a second subset of the features associated with a negative feedback; and
modifying, by the learning module, the tunable parameters based on the first and second subsets of the features.

14. The method of claim 13 , wherein modifying the tunable parameters ameliorates correlation with the second subset of the features, and produces correlation with the first subset of the features.

15. The method of claim 13 , further comprising using data stored to represent the at least one of the stored electronically-represented projects comprising the first subset of the features to further develop and manage the received electronically-represented project.

16. The method of claim 11 , further comprising invoking the learning module repeatedly to compute the adjustments.

17. A non-transitory computer-readable storage medium storing machine readable instructions that upon execution by at least one processor cause the at least one processor to:
compare, by a project-comparator subsystem executed by the at least one processor, the received electronically-represented project to at least one of stored electronically-represented projects, wherein to compare comprises:
extract features from the received electronically-represented project and the at least one of the stored electronically-represented projects;
determine a similarity metric for each of the extracted features; and
determine an overall similarity metric by applying weights to the similarity metric for the extracted features, and computationally combining the weighted similarity metrics; and

receive a feedback regarding similarity between the received electronically-represented project and the at least one of the stored electronically-represented projects; and
compute adjustments to the project-comparator subsystem in response to the received feedback.

18. The non-trans

Assignees

• Entit Software Llc

Inventors

• Hickey Marianne
• Trastour David
• Byde Andrew
• Rahmouni Maher
• Bartolini Claudio

Classifications

• G06Q10/06

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

• G06N99/005

  Physics · mapped topic

• G06K9/00483

  Physics · mapped topic

• G06F17/30011 Primary

  Physics · mapped topic

• G06V30/418

  Document matching, e.g. of document images · CPC title