Systems and methods of natural language processing to rank users of real time communications connections
Inventors
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-9819802-B2
Publication Date
2017-11-14
Expiration Date
Abstract
A computing apparatus configured to perform natural language processing, e.g., by comparing the words in a communication transcript of a user to a list of keywords, to generate an input vector representing a pattern of text in the communication transcript. A predictive model is generated from correlating input vectors to user ranking scores, e.g., for retention. The input vector determined from a communication transcript is applied to the computation model to compute a predicted retention score of the user. The retention score can be used, for example, to select a personalized recommendation for a communication connection to an adviser and/or a targeted offer.
Core Innovation
The invention provides a connection server configured to establish real time communication connections between telephonic devices. The connection server is coupled with a data store storing communication references of a first set of telephonic devices registered in the data store and information about first users of the first set of telephonic devices. In response to requests received from second users of a second set of telephonic devices identifying selected ones of the first users, the connection server establishes communication connections between respective telephonic devices of the second users and respective telephonic devices of the selected ones of the first users based on communication references in the data store. The system generates transcripts of communications over the communication connections and stores communication records of the communication connections in the data store.
A computing apparatus performs natural language processing on the communications transcripts to generate input vectors. The computing apparatus generates ranking scores of the second users from the communication records, where the input vectors are correlated with the ranking scores to establish a predictive model of ranking score. After the connection server provides a communication connection between a telephonic device of a user and a telephonic device of selected ones of the first users, the computing apparatus generates a transcript of a communication between the user and the selected ones of the first users. The computing apparatus performs natural language processing on the transcript to generate first input vectors and applies the first input vectors in the predictive model of ranking score to generate a ranking score of the user.
The disclosed approach correlates communications-derived natural language processing features with user ranking scores to train and use a predictive model of ranking score. The invention further refines the natural language processing and predictive model by using a keyword list and term frequencies in the transcripts to form input vectors, and by using statistical correlation analysis and/or a machine learning technique to establish the predictive model. The invention can also establish the communication connection by establishing first and second telephonic-device connections and bridging them, and can generate transcripts using voice recognition as part of the natural language processing workflow.
Claims Coverage
The independent claims (clm-00001, clm-00012, clm-00013) consistently cover five inventive features: a real time telephonic connection framework, transcript generation and storage, natural language processing to generate input vectors, correlation of input vectors with ranking scores to establish a predictive model, and application of that model to generate a user ranking score.
Real time communication connections between telephonic devices using stored communication references
A connection server configured to establish real time communication connections between telephonic devices, with the connection server coupled with a data store storing communication references of a first set of telephonic devices and information about first users; wherein the connection server establishes communication connections in response to requests that identify selected ones of the first users based on the stored communication references.
Transcripts and communication records from established communication connections
Generating transcripts of communications over the communication connections established between second users and selected ones of the first users and storing communication records of the communication connections in the data store.
Natural language processing to generate input vectors from transcripts
Performing natural language processing on the transcripts to generate input vectors, including generating a transcript of a communication between a user and the selected ones of the first users and performing natural language processing on that transcript to generate first input vectors.
Correlating input vectors with ranking scores to establish a predictive model of ranking score
Generating ranking scores of the second users from the communication records, wherein the input vectors are correlated with the ranking scores to establish a predictive model of ranking score.
Applying the predictive model of ranking score to generate a user ranking score
Applying the first input vectors in the predictive model of ranking score to generate a ranking score of the user after the connection server provides a communication connection between the user and the selected ones of the first users.
The independent claims collectively require a connection server that establishes real time communication connections based on stored communication references, transcript generation and storage of communication records, NLP to generate input vectors, correlation of input vectors with ranking scores to establish a predictive model, and application of the predictive model to generate a ranking score for a user based on transcript-derived first input vectors.
Stated Advantages
Documented Applications
No documented applications found
Interested in licensing this patent?