Determining trustworthiness and compatibility of a person

• US 9,070,088 B1
• Filed: 09/16/2014
• Issued: 06/30/2015
• Est. Priority Date: 09/16/2014
• Status: Active Grant

• Current Assignee: AirBNB Incorporated
• Original Assignee: Trooly, Inc. (AirBNB Incorporated)
• Inventors: Baveja, Sarabjit Singh, Sarma, Anish Das, Dalvi, Nilesh
• Primary Examiner(s): Mamillapalli, Pavan
• Application Number:
  US14/488,177
• Time in Patent Office: 287 Days
• Field of Search: 706/14
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 21/316 by observing the pattern of computer usage, e.g. typical user behaviour
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06N 5/048 Fuzzy inferencing

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for identifying documents related to a person, deriving behavior and personality trait metrics from analyzing the documents for information relevant to assessing behavior and personality of the person, and determine a trustworthiness score or compatibility score of the person based on the behavior and personality trait metrics using a scoring system.

Claims

1. A method for providing a trustworthiness score of a person comprising:

obtaining a request for a trustworthiness score of a person from a requestor;

for a plurality of different behavior and personality traits, identifying respective portions of a respective plurality of documents that contain information relevant to predicting the trait based on a type of the document indicated by one or more of a source of the document, a layout of the document, a content type of the document, and a data or metadata contained in the document;

deriving a behavior trait metric for each behavior trait by analyzing the respective document portions for the behavior trait;

deriving a personality trait metric for each personality trait by analyzing the respective document portions for the personality trait;

providing the behavior trait metrics and the personality trait metrics as input to a scoring system and obtaining as output from the system a trustworthiness score of the person, wherein the scoring system is rule based and a machine learning system;

providing the trustworthiness score of the person to the requestor; and

wherein obtaining, identifying, deriving and providing are performed by one or more computer processors.

2. The method of claim 1, further comprising:

determining that trustworthiness score is not accurate; and

retraining the scoring system to correct for the inaccuracy using the behavior trait metrics and the personality trait metrics.

3. The method of claim 2 wherein determining that the trustworthiness score is not accurate comprises receiving an indication of inaccuracy from the requestor, the person, a third party, a trained classifier, or a rule-based system.

4. The method of claim 1 wherein the trustworthiness score is based on personality and behavior traits that predict the likelihood of the person being a positive actor in an online or offline person-to-person interaction.

5. The method of claim 1 wherein deriving a particular trait metric comprises:

calculating a respective identity score for each of the documents wherein the identity score is based on one or more identification attributes of the person that match identity information in the document;

identifying one or more content attributes in the identified portions for the trait in each of the documents that occur in a respective dictionary or directory for the particular trait and calculating a respective initial score for the document based on, at least, weights associated with the identified content attributes and the identity score of the document; and

combining the initial scores to calculate the particular trait metric for the person.

6. The method of claim 5, wherein a content attribute is a word, phrase, image, tag, header, video, link, symbol, number, or connection to another individual or webpage.

7. The method of claim 5, wherein weights are determined based on rule based or machine based learning of the strength of the relationship between each content attribute and each trait.

8. The method of claim 5 wherein identifying content attributes in a particular document comprises identifying one or more portions of the document that contain text authored by the person or that provide information about the person.

9. The method of claim 8 wherein a particular identified portion indicates that the person created a false or misleading online profile or provided false or misleading information to a service provider, is involved with drugs or alcohol, is involved with hate websites or organizations, is involved in sex work, perpetrated a crime, is involved in civil litigation, is a known fraudster or scammer, is involved in pornography, has authored online content with negative language, or has interests that indicate negative personality or behavior traits.

10. The method of claim 1, further comprising:

verifying that the requestor is authorized to receive the trustworthiness score of the person.

11. The method of claim 1, further comprising:

adding the trustworthiness score for the person to a database of trustworthiness scores for a plurality of different people.

12. The method of claim 1, further comprising:

receiving an indication that a second person trusts the person;

creating a relationship between the second person and the person; and

adjusting the trustworthiness score of the person based on, at least, the trustworthiness score of the second person.

13. The method of claim 1, further comprising:

obtaining identification attributes of the person;

determining whether the identification attributes are authentic;

deriving the one or more behavior and personality trait metrics if the identification attributes are authentic; and

deriving additional identification attributes and iteratively deriving additional behavior and personality trait metrics based on the additional identification attributes.

14. The method of claim 13 wherein determining whether the identification attributes are authentic comprises:

determining if the person is associated with a fake social network profile.

15. The method of claim 13 wherein determining whether the identification attributes are authentic comprises:

determining if identity information about the person contained in one or more documents is consistent with the identification attributes.

16. The method of claim 13, further comprising:

extracting identity information for the person from one or more documents;

verifying that the identity information is authentic; and

adding the identity information to the identification attributes.

17. The method of claim 1, wherein a particular personality trait is badness, anti-social tendencies, goodness, conscientiousness, openness, extraversion, agreeableness, neuroticism, narcissism, Machiavellianism, or psychopathy.

18. The method of claim 1, wherein a particular behavior trait is creating a false or misleading online profile or providing false or misleading information to a service provider, involvement with drugs or alcohol, involvement with hate websites or organizations, involvement in sex work, involvement in a crime, involvement in civil litigation, being a known fraudster or scammer, involvement in pornography, or authoring online content with negative language.

19. The method of claim 5 wherein the identification attributes of the person comprise a plurality of the following:

name, email address, telephone number, geographic location, date of birth, social connections, employment history, education history, driver'"'"'s license number, financial account information, Internet Protocol (IP) address, and device identifier.

20. The method of claim 1, further comprising:

obtaining identification attributes of the person; and

identifying the documents including;

formulating a plurality of queries using the personal identification attributes, the queries configured to maximize retrieval of relevant documents and information;

submitting the queries to one or more remote systems; and

receiving the search results from the remote systems in response to submitting the queries wherein each search result identifies a respective document of the documents.

21. The method of claim 1 wherein a particular document is a web page, information from a database, a post on the person'"'"'s social network account, a post on a blog or a microblog account of the person, a comment made by the person on a website, or a directory listing for a company or association.

22. A memory for storing data for access by a computer program being executed on one or more computer processes, the memory comprising:

a data structure containing information stored in the memory for use by the computer program, the data structure comprising;

a plurality of person data objects stored in the memory, each of the person data objects comprising a trustworthiness score, wherein the scoring system is rule based and a machine learning system, one or more behavior trait metrics, one or more personality trait metrics, and one or more compatibility scores between the person and one or more other persons, each behavior or personality trait metric having been derived by analyzing respective portions of documents that contain information relevant to predicting a corresponding behavior or personality trait based on a type of the document indicated by one or more of a source of the document, a layout of the document, a content type of the document, and a data or metadata contained in the document; and

a plurality of relationships between a plurality of the person data objects wherein each of the relationships comprises a link between a first user and a second user, and a weight of the link.

23. The memory of claim 22, wherein a particular personality trait is badness, anti-social tendencies, goodness, conscientiousness, openness, extraversion, agreeableness, neuroticism, narcissism, Machiavellianism, or psychopathy.

24. The memory of claim 22, wherein a particular behavior trait is creating a false or misleading online profile or providing false or misleading information to a service provider, involvement with drugs or alcohol, involvement with hate websites or organizations, involvement in sex work, involvement in a crime, involvement in civil litigation, being a known fraudster or scammer, involvement in pornography, or authoring online content with negative language.

25. The memory of claim 22, wherein a particular person data object further comprises identification attributes of a person represented by the particular person data object.

26. The memory of claim 25, wherein the identification attributes comprise a plurality of the following:

name, email address, telephone number, geographic location, date of birth, social connections, employment history, education history, driver'"'"'s license number, financial account information, IP address, and device identifier.

27. The memory of claim 22, wherein the trustworthiness score is based on personality and behavior traits that predict the likelihood of the person being a positive actor in an online or offline person-to-person interaction.

28. The memory of claim 22, wherein each relationship further comprises a calculated compatibility between the first user and the second user based on respective behavior and personality trait metrics of the first user and the second user.

29. The memory of claim 22, wherein each relationship is determined by one or more online interactions between the first user and the second user.

30. A method for providing a compatibility score of a person comprising:

obtaining a request for a compatibility score between two persons from a requestor;

for a plurality of different behavior and personality traits, identifying respective portions of a respective plurality of documents that contain information relevant to predicting the trait of a first person based on a type of the document indicated by one or more of a source of the document, a layout of the document, a content type of the document and a data or metadata contained in the document;

deriving a behavior trait metric of the first person for each behavior trait by analyzing the respective document portions for the behavior trait;

deriving a personality trait metric of the first person for each personality trait by analyzing the respective document portions for the personality trait;

providing the behavior trait metrics and the personality trait metrics of the first person and corresponding metrics for a second person as input to a scoring system and obtaining as output from the system a compatibility score between the two persons;

providing the compatibility score to the requestor; and

wherein obtaining, identifying, deriving and providing are performed by one or more computer processors.