Methods and systems for creating a classifier capable of predicting personality type of users

Patent metadata

Field	Value
Publication number	US-10013659-B2
Application number	US-201514731445-A
Country	US
Kind code	B2
Filing date	Jun 5, 2015
Priority date	Nov 7, 2014
Publication date	Jul 3, 2018
Grant date	Jul 3, 2018

Abstract

The disclosed embodiments illustrate methods and systems for creating a classifier for predicting a personality type of users. The method includes receiving a first tag for messages, from a crowdsourcing platform. The first tag relates to personality type of users. Further, the messages, tagged with first tag are segregated into a training data and a testing data. Further, parameters associated with set of messages in the training data are determined based on type of messages. Further, classifiers are trained for a personality type. Further, a second tag for set of messages in testing data is predicted using trained classifiers for a combination of parameters. A performance of classifiers is determined by comparing the second tag and the first tag associated with set of messages in the testing data. A classifier is selected from classifiers, which is indicative of a best combination of parameters to predict personality type of users.

First claim

What is claimed is:

1. A method for predicting a personality type of one or more users, the method comprising:
retrieving, by one or more microprocessors, one or more messages from a social media platform, wherein the one or more messages include at least one of an audio message, a video message, or a text message;
receiving, by a transceiver, a first tag associated with each of the one or more messages, from a crowdsourcing platform, wherein the first tag relates to the personality type of the one or more users;
segregating, by the one or more microprocessors, the one or more messages, tagged with the first tag, into a first data set and a second data set, wherein the first data set corresponds to a training data, wherein the second data set corresponds to a testing data, and wherein each of the training data and the testing data comprises a set of messages from the one or more messages;
determining, by the one or more microprocessors, one or more parameters associated with the set of messages in the training data based on the one or more messages;
determining, by the one or more microprocessors, one or more combinations of the one or more parameters;
training, by the one or more microprocessors, one or more classifiers for the personality type, wherein each of the one or more classifiers is trained for a combination from the one or more combinations of the one or more parameters;
predicting, by the one or more microprocessors, a second tag for the set of messages in the testing data using the trained one or more classifiers;
determining, by the one or more microprocessors, a performance of each of the one or more classifiers, based on a comparison of the second tag with the first tag associated with the set of messages in the testing data;
ranking, by the one or more microprocessors, the one or more classifiers based on the performance of each of the one or more classifiers;
selecting, by the one or more microprocessors, a classifier from the one or more classifiers based on a result of the ranking, wherein the selected classifier is indicative of a best combination from the one or more combinations of the one or more parameters to predict the personality type of the one or more users;
using, by the one or more microprocessors, the selected classifier to categorize at least one of another audio, video, and text messages to identify target users of a product or service; and
transmitting, by the transceiver, the identified target users to one or more organizations for the one or more organizations to place advertisement of the product or service on the social media platform targeting the identified target users.

2. The method of claim 1 , wherein the one or more parameters associated with the set of messages are determined by using at least one of signal processing techniques, image processing techniques, or natural language processing techniques.

3. The method of claim 1 , wherein the one or more parameters associated with the audio message comprise at least one of a duration of the audio message, a duration for which a user speaks in the audio message, a rate of speech of the user, a pitch of the user, or a number of pauses taken by the user.

4. The method of claim 1 , wherein the one or more parameters associated with the video message comprise at least one of a posing/body language of a user, or movement parameters of the user, wherein the posing/body language of the user comprises at least one of a duration for which the user looks at a camera, a number of instances in which the user looked away from the camera, a proximity of the user to the camera, wherein the movement parameters of the user comprise at least a degree of excitement of the user.

5. The method of claim 1 , wherein the one or more parameters associated with the text message comprise at least one of word tokens, word sentiments, or word statistical parameters.

6. The method of claim 1 , wherein the personality type comprises at least one of an openness to experience, a conscientiousness, an extraversion, an agreeableness, or a neuroticism.

7. The method of claim 1 , wherein the social media platform comprises at least one of social networking websites, chat/messaging applications, web-blogs, online communities, web-forums, community portals, or online interest groups.

8. The method of claim 1 , wherein the one or more classifiers are trained using one or more regression techniques comprising at least one of a ridge regression technique, a logistic regression technique, or a binary classification technique.

9. The method of claim 1 , wherein the one or more combinations of the one or more parameters comprise at least one of audio-video parameters, a combination of the audio-video parameters and word statistical parameters, a combination of the word statistical parameters and word sentiments, or a combination of the word statistical parameters, the word sentiments, and a gender of a user.

10. The method of claim 1 , wherein the performance of each of the one or more classifiers is determined based on one or more performance parameters comprising at least one of a precision, a recall, and an f-measure.

11. The method of claim 1 further comprising ranking, by the one or more microprocessors, the one or more classifiers for the personality type based on the performance.

12. The method of claim 1 further comprising determining, by the one or more microprocessors, a score of each of the one or more messages against the personality type based on the first tag associated with each of the one or more messages, wherein the score is indicative of a positive score, or a negative score.

13. A system predicting a personality type of one or more users, the system comprising:
a transceiver is configured to receive a first tag associated with each of one or more messages, from a crowdsourcing platform, wherein the first tag relates to the personality type of the one or more users, wherein the one or more messages are shared by the one or more users on one or more social media platforms, wherein the one or more messages include at least one of an audio message, a video message, or a text message;
one or more microprocessors configured to:
segregate the one or more messages, tagged with the first tag, into a first data set and a second data set, wherein the first data set corresponds to a training data, wherein the second data set corresponds to a testing data, and wherein each of the training data and the testing data comprises a set of messages from the one or more messages;
determine one or more parameters associated with the set of messages in the training data based on the one or more types of messages;
determine one or more combinations of the one or more parameters;
train one or more classifiers for the personality type, wherein each of the one or more classifiers is trained for a combination from the one or more combinations of the one or more parameters;
predict a second tag for the set of messages in the testing data using the trained one or more classifiers;
determine a performance of each of the one or more classifiers, based on a comparison of the second tag with the first tag associated with the set of messages in the testing data;
rank the one or more classifiers based on the performance of each of the one or more classifiers;
select a classifier from the one or more classifiers based on a result of the ranking, wherein the selected classifier is indicative of a best combination from the one or more combinations of the one or more parameters to predict the personality type of the one or more users; and
use the selected classifier to categorize at least one of another audio, vi

Assignees

• Xerox Corp
• Conduent Business Services Llc

Inventors

• Li Juan
• Bhatia Sumit
• Sarkar Chandrima

Classifications

• G06Q10/40

  Business processes related to social networking or social networking services · CPC title

• G06Q50/01

  Physics · mapped topic

• G06N99/005 Primary

  Physics · mapped topic

• G06N20/00 Primary

  Machine learning · CPC title