System and methods for managing communications events

Abstract

An example communications routing system includes a prediction engine to, for each user account, generate a predicted cost of communications events using a gamma-gamma cost model, generate a predicted revenue of presented advertisements using a gamma-gamma revenue model, generate a margin based on the predicted cost and the predicted revenue, and, when the margin for a given user account is below a margin threshold, restrict a status associated with the given user account.

What is claimed is: 
     
         1 . A communications routing system comprising:
 a routing engine configured to manage Voice over Internet Protocol (VoIP) communications events between a plurality of communications devices using respective VoIP telephone numbers;   an account database storing a plurality of unique associations between user accounts of users and the VoIP telephone numbers;   a cost database storing a cost of the communications events for the user accounts;   a revenue database storing a revenue of presented advertisements associated with the user accounts;   a prediction engine configured to, for each of the user accounts:
 generate a predicted cost of the communications events using a gamma-gamma cost model; 
 generate a predicted revenue of the presented advertisements using a gamma-gamma revenue model; 
 generate a margin based on the predicted cost and the predicted revenue; and 
 when the margin for a given user account is below a margin threshold, restrict a status associated with the given user account. 
   
     
     
         2 . The communications routing system of  claim 1 , wherein, to generate the predicted revenue for the given user account, the prediction engine is configured to:
 determine optimal parameters of the gamma-gamma revenue model based on a maximum likelihood estimation; and   determine a predictive revenue distribution for the given user account using an inverse gamma revenue distribution based on (i) the optimal parameters, (ii) an average daily advertisement revenue for the given user account and (iii) a number of days on which the given user account was presented with an advertisement.   
     
     
         3 . The communications routing system of  claim 1 , wherein, to generate the predicted cost for the given user account, the prediction engine is configured to:
 determine optimal parameters of the gamma-gamma cost model based on a maximum likelihood estimation; and   determine a predictive cost distribution for the given user using an inverse gamma distribution based on (i) the optimal parameters, (ii) an average daily cost of communications events for the given user account and (iii) a number of days on which a communications event occurs for given user account.   
     
     
         4 . The communications routing system of  claim 1 , wherein, in response to restricting the status of the given user account the routing engine is configured to:
 block communications events for the given user account;   limit features of communications services available for the given user account;   increase a length of advertisements presented to the given user account; or   increase a number of advertisements presented to the given user account.   
     
     
         5 . The communications routing system of  claim 1 , wherein the prediction engine is further configured to:
 determine an upper revenue confidence bound representing a revenue threshold percentile of an inverse gamma revenue distribution;   determine a lower cost confidence bound representing a cost threshold percentile of an inverse gamma cost distribution; and   generate the margin based on a difference between the upper revenue confidence bound and the lower cost confidence bound.   
     
     
         6 . The communications routing system of  claim 5 , wherein the revenue threshold percentile is a 95 th  percentile. 
     
     
         7 . The communications routing system of  claim 5 , wherein the cost threshold percentile is a 5 th  percentile. 
     
     
         8 . The communications routing system of  claim 1 , wherein the prediction engine is configured to update the account database to include an indication of the restricted status of the given user account. 
     
     
         9 . A method comprising:
 storing a plurality of unique associations between user accounts of users and VoIP telephone numbers;   storing a cost of communications events for the user accounts;   storing a revenue of presented advertisements associated with the user accounts;   for each of the user accounts:
 generating a predicted cost of the communications events using a gamma-gamma cost model; 
 generating a predicted revenue of the presented advertisements using a gamma-gamma revenue model; 
 generating a margin based on the predicted cost and the predicted revenue; and 
 when the margin for a given user account is below a margin threshold, restricting a status associated with the given user account. 
   
     
     
         10 . The method of  claim 9 , wherein generating the predicted revenue for the given user account comprises:
 determining optimal parameters of the gamma-gamma revenue model based on a maximum likelihood estimation; and   determining a predictive revenue distribution for the given user account using an inverse gamma revenue distribution based on (i) the optimal parameters, (ii) an average daily advertisement revenue for the given user account and (iii) a number of days on which the given user account was presented with an advertisement.   
     
     
         11 . The method of  claim 9 , wherein generating the predicted cost for the given user account comprises:
 determining optimal parameters of the gamma-gamma cost model based on a maximum likelihood estimation; and   determining a predictive cost distribution for the given user using an inverse gamma distribution based on (i) the optimal parameters, (ii) an average daily cost of communications events for the given user account and (iii) a number of days on which a communications event occurs for given user account.   
     
     
         12 . The method of  claim 9 , further comprising, in response to restricting the status of the given user account:
 blocking communications events for the given user account;   limiting features of communications services available for the given user account;   increasing a length of advertisements presented to the given user account; or   increasing a number of advertisements presented to the given user account.   
     
     
         13 . The method of  claim 9 , further comprising:
 determining an upper revenue confidence bound representing a revenue threshold percentile of an inverse gamma revenue distribution;   determining a lower cost confidence bound representing a cost threshold percentile of an inverse gamma cost distribution; and   generating the margin based on a difference between the upper revenue confidence bound and the lower cost confidence bound.   
     
     
         14 . The method of  claim 13 , wherein the revenue threshold percentile is a 95 th  percentile. 
     
     
         15 . The method of  claim 13 , wherein the cost threshold percentile is a 5 th  percentile. 
     
     
         16 . The method of  claim 9 , further comprising updating an account database to include an indication of the restricted status of the given user account. 
     
     
         17 . A communications system comprising:
 a mobile network to provide mobile communications services for a plurality of mobile communications devices over a wide-area IP network;   a wireless local-area network (WLAN) to provide wireless network communications services for a plurality of communications devices over the wide-area IP network;
 a Voice over Internet Protocol (VoIP) service coupled to the mobile network and the WLAN via the wide-area IP network to provide VoIP communications services for the plurality of communications devices; 
 a public switched telephone network (PSTN) coupled to the VoIP service; 
   a communications routing system coupled to the plurality of communications devices via the wide-area IP network to manage Voice over Internet Protocol (VoIP) communications events between a plurality of communications devices using respective VoIP telephone numbers;
 an account database storing a plurality of unique associations between user accounts of users and the VoIP telephone numbers; 
 a cost database storing a cost of the communications events for the user accounts; 
 a revenue database storing a revenue of presented advertisements associated with the user accounts; 
 a prediction engine configured to, for each of the user accounts:
 generate a predicted cost of the communications events using a gamma-gamma cost model; 
 generate a predicted revenue of the presented advertisements using a gamma-gamma revenue model; 
 generate a margin based on the predicted cost and the predicted revenue; and 
 
   when the margin for a given user account is below a margin threshold, restrict a status associated with the given user account.