Generating optimal strategy for providing offers

Abstract

Generating optimal strategies for providing offers to a plurality of customers is described. A plurality of categorical attributes (for example, gender and residential status) and ordinal attributes (for example, risk score and credit line utilization) can be determined. Values of one of more categorical attributes can be changed as per a transition probability table. Some probabilities can be varied to determine a first tradeoff, based on which a first updated strategy can be generated. Further, noise can be added to one or more ordinal attributes. Standard deviation of a noise distribution associated with the noise can be varied so as to determine a second tradeoff, based on which a second updated strategy can be generated. The second updated strategy can be an update of the first updated strategy. Offers can be provided to the plurality of customers in accordance with the second updated strategy.

1 . A method comprising:
 obtaining, by at least one data processor, data associated with a plurality of individuals, each individual associated with a plurality of attributes;   determining, by at least one data processor and using the obtained data, a best offer for each attribute;   forming, by at least one data processor, a decision tree characterizing best offers for corresponding attributes;   comparing, by at least one data processor, performance of the decision tree with performance of a challenger decision tree to obtain a best performing decision tree;   providing, by at least one data processor, offers to the plurality of individuals in accordance with the best performing decision tree.   
     
     
         2 . The method of  claim 1 , further comprising:
 determining, by at least one data processor and based on the obtained data, a plurality of possible offers for the plurality of individuals,   wherein the best offers for each attribute are selected from the plurality of possible offers.   
     
     
         3 . The method of  claim 1 , further comprising:
 determining, by at least one data processor and using the obtained data, a plurality of causal models; and   forming, by at least one data processor and by joining two or more causal models, a decision model that evaluates one or more objectives of an entity, the causal model being used to determine the best offer for each attribute.   
     
     
         4 . The method of  claim 3 , wherein the causal models characterize a response of an individual to a historical offer. 
     
     
         5 . The method of  claim 3 , wherein the determining of the best offer is based on evaluation of at least one of a global maximum value and a local maximum value by the decision model. 
     
     
         6 . The method of  claim 1 , wherein the challenger decision trees are obtained by changing values of some attributes associated with the decision tree. 
     
     
         7 . The method of  claim 1 , wherein:
 the performance of the decision tree is characterized by business efficacy provided by implementing a strategy associated with the decision tree, the business efficacy associated with the decision tree being based on a plurality of iterations of strategy evolution; and   the performance of the challenger decision tree is characterized by business efficacy provided by implementing a strategy associated with the challenger decision tree, the business efficacy associated with the challenger decision tree being based on a plurality of iterations of strategy evolution.   
     
     
         8 . The method of  claim 7 , wherein the business efficacy characterizes a profit of an entity providing the offers to the plurality of individuals. 
     
     
         9 . A method comprising:
 obtaining, by at least one data processor, a graph characterizing a strategy for providing offers;   modifying, by at least one data processor, one or more attributes associated with a plurality of individuals, the one or attributes represented by the graph;   generating, by at least one data processor and based on the modified one or more attributes, an updated strategy for providing offers.   
     
     
         10 . The method of  claim 9 , wherein the graph includes a plurality of dots having at least one of a corresponding color and a corresponding intensity, the at least one of the corresponding color and the corresponding intensity characterizing a value of at least one attribute for an associated individual. 
     
     
         11 . The method of  claim 9 , wherein the offers are provided to the plurality of individuals. 
     
     
         12 . The method of  claim 9 , wherein the modifying of the one or more attributes comprises adding noise to the attributes. 
     
     
         13 . The method of  claim 12 , wherein the adding of the noise to the attributes comprises varying a standard deviation of a noise distribution to determine the noise, the adding of the noise to the attributes providing an optimal profit to an entity providing the offers, the optimal profit being more than a profit obtained without the addition of the noise. 
     
     
         14 . A non-transitory computer program product storing instructions that, when executed by at least one programmable processor, cause the at least one programmable processor to perform operations comprising:
 determining attributes associated with a strategy;   adding Gaussian noise to one or more attributes;   varying standard deviation of the Gaussian noise to determine a tradeoff; and   generating an updated strategy associated with the tradeoff.   
     
     
         15 . The computer program product of  claim 14 , wherein the at least one programmable processor further performs operations comprising:
 providing offers based on the updated strategy.   
     
     
         16 . The computer program product of  claim 14 , wherein the updated strategy is determined based on the tradeoff, the tradeoff characterizing a balance between cost of a business entity and rate of update of strategies. 
     
     
         17 . A method comprising:
 determining, by at least one data processor, a plurality of attributes associated with a strategy;   changing, by at least one data processor, values of a first set of one or more attributes in accordance with a transition probability table;   varying, by at least one data processor, one or more probabilities to determine a first tradeoff; and   generating, by at least one data processor and based on the first tradeoff, a first updated strategy.   
     
     
         18 . The method of  claim 17 , further comprising:
 adding, by at least one data processor, noise to a second set of one or more attributes;   varying, by at least one data processor, standard deviation of a noise distribution associated with the noise to determine a second tradeoff; and   generating, by at least one data processor and based on the second tradeoff, a second updated strategy, the second updated strategy characterizing an update of the first updated strategy.   
     
     
         19 . The method of  claim 18 , further comprising:
 providing, by at least one data processor and based on the second updated strategy, offers to a plurality of individuals.   
     
     
         20 . The method of  claim 18 , wherein the first set of one or more attributes include gender and residential status. 
     
     
         21 . The method of  claim 20 , wherein the second set of one or more attributes include risk score and credit line utilization. 
     
     
         22 . The method of  claim 18 , further comprising:
 determining, from a table, eligibility constraints for provision of one or more offers to one or more customers,   wherein at least one of the first updated strategy and the second updated strategy are based on the eligibility constraints to exclude provision of some offers to corresponding ineligible customers.   
     
     
         23 . The method of  claim 17 , wherein at least one of the first tradeoff and the second tradeoff are determined using corresponding tradeoff curves, each tradeoff being characterized by a sweet-spot on a corresponding tradeoff curve, the sweet-spot characterizing a position where generated strategy data is more than a first threshold while profit is more than a second threshold.