System and methods for customer quality prediction

1 . A process to determine customer value, the process comprising:

determining, via a customer ranking engine (CRE), comprising a machine learning or artificial intelligence system comprising program instructions, stored in a memory, configured to execute on one or more processors in the CRE, wherein the memory is in electrical communication with the one or more processors and comprises a program memory and a data memory, scores rating historical customer quality, wherein each score is determined from a plurality of multi-source data components including at least credit score, income history, and maintenance request data, and wherein each score is a vector quantity determined as a resultant of the multi-source data components;

determining, via the one or more processors, a customer quality threshold, wherein the customer quality threshold is based at least in part on quality scores and market segment preferences and indications received from the user via a user interface;

retrieving, via the one or more processors, historical customer data from multiple sources;

dividing, via the one or more processors, historical customer data into at least a train partition and test partition;

rating, via the one or more processors, a plurality of customers based on a quality score, wherein said quality score is based at least in part on a function of historical training data associated with said plurality of customers;

determining, via the one or more processors, a subset of customers from said plurality of customers, wherein said subset of customers comprises one or more customers that do not meet the customer quality threshold;

training, via the one or more processors, a neural network, to make determinations as to whether one or more customers of the plurality of customers satisfy the customer quality threshold, wherein said training of the neural network is based at least in part on cross-validation with the test partition, wherein training comprises iteratively adjusting weights of the neural network using the train partition until a model prediction error of the neural network satisfies a predetermined model prediction error threshold, and wherein the model prediction error threshold is determined based on the cross-validation with the test partition;

testing, via the one or more processors, the neural network, to recognize historical customer quality determined as a function of ranking the quality scores, and determining, based on the cross-validation, whether the neural network is overfit by evaluating a function of error in the cross-validation performed by the processor, and upon a determination that the neural network is overfit, discarding the neural network;

upon a determination that the neural network is not overfit, deploying the trained neural network for production prediction operation;

determining, via the one or more processors, a customer quality improvement rate (CQIR) for one or more customers of the subset of customers, and a CQIR threshold, wherein the CQIR is a slope of a quality prediction line through at least two points on a quality prediction determined by the processor based on customer data and an output of the trained neural network, the slope defining a rate at which a customer quality changes over time and the CQIR threshold is a rate of change of CQIR required to determine whether an individual customer meets desired potential customer quality trends, wherein the CQIR and CQIR threshold are based at least in part on use of said trained neural network;

predicting future customer quality for a first set of customers from said subset of customers based at least in part on said CQIR and said CQIR threshold and said trained neural network; and

rejecting said first set of customers from said plurality of customers whose CQIR does not satisfy the CQIR threshold, thereby automatically filtering customers from a pool of potential customers based on the deployed trained neural network.

2 . The process of claim 1 , wherein the scores further comprise multi-source data.

3 . The process of claim 1 , wherein customer quality further comprises a vector quantity.

4 . The process of claim 1 , wherein rating historical customer quality further comprises each score determined as a combination of multiple data components.

5 . The process of claim 1 , wherein training the neural network further comprises partitioning historical data into train and test partitions.

6 . The process of claim 1 , wherein training the neural network further comprises cross validation.

7 . The process of claim 1 , wherein training the neural network further comprises a test to determine if the model prediction error is less than a predetermined maximum.

8 . The process of claim 1 , wherein training the neural network further comprises a test to determine if the neural network is overfit based on cross-validation.

9 . The process of claim 1 , wherein determining scores further comprises normalizing data.

10 . A process to determine customer value, the process comprising:

determining, via a customer ranking engine (CRE), comprising a machine learning or artificial intelligence system comprising program instructions, stored in a memory, configured to execute on one or more processors in the CRE, wherein the memory is in electrical communication with the one or more processors and comprises a program memory and a data memory, scores rating historical customer quality, wherein the scores are determined as a function of multi-source data, and wherein customer quality is a vector quantity determined as a function of a plurality of multi-source data components;

training, via the one or more processors, a neural network, to recognize historical customer quality determined as a function of ranking the scores, wherein training the neural network includes partitioning historical data into train and test partitions, and, training the neural network using the train partition until the model prediction error of the neural network satisfies a predetermined threshold, wherein training comprises iteratively adjusting weights of the neural network and evaluating the model prediction error after each iteration, and upon a determination that the model prediction error did increase, discarding the neural network;

testing, via the one or more processors, the predictive analytic model, via the neural network, wherein testing the neural network includes cross-validation based on the test partition, and determining whether the neural network is overfit based on a function of error in the cross-validation, and upon a determination that the neural network is not overfit, deploying the trained neural network for production prediction operation;

determining, via the one or more processors, a customer quality improvement rate (CQIR) and a CQIR threshold, wherein the CQIR is a slope of a quality prediction line through at least two points determined by the processor based on customer data and the neural network, the slope defining a rate at which customer quality changes over time and the CQIR threshold is a rate of change of CQIR required to determine whether an individual customer meets desired potential customer quality trends, wherein the CQIR and CQIR threshold are based at least in part on use of said trained neural network; and,

predicting, via the one or more processors, future customer quality based at least in part on the CQIR and said CQIR threshold and the trained neural network, and automatically rejecting customers whose CQIR does not satisfy the CQIR threshold.

11 . The process of claim 10 , wherein the neural network further comprises a random forest ensemble layer integrated with the neural network.

12 . The process of claim 10 , wherein the neural network further comprises a decision tree component integrated with the neural network.

13 . The process of claim 10 , wherein the neural network further comprises a Bayesian classifier component integrated with the neural network.

14 . The process of claim 10 , wherein the train partition further comprises labeled data.

15 . The process of claim 10 , wherein ranking customer quality further comprises evaluating an angle determined by vector analysis of data components.

16 . A process to determine customer value, the process comprising:

determining, via a customer ranking engine (CRE), comprising a machine learning or artificial intelligence system comprising program instructions, stored in a memory, configured to execute on one or more processors in the CRE, wherein the memory is in electrical communication with the one or more processors and comprises a program memory and a data memory, scores rating historical customer quality, wherein the scores are determined as a function of multi-source data components, wherein customer quality is a vector quantity determined as a function of a plurality of the multi-source data components;

training, via the one or more processors, a neural network to recognize historical customer quality determined as a function of ranking the scores, wherein training the neural network includes partitioning historical data into train and test partitions, and training the predictive analytic model using the train partition until the model prediction error of the neural network satisfies a predetermined threshold, wherein training comprises iteratively adjusting weights of the neural network and evaluating the model prediction error after each iteration, and wherein ranking the scores further comprises evaluating an angle determined by vector analysis of data components;

testing the neural network, wherein testing the model includes cross-validation based on the test partition, and determining whether the neural network is overfit based on a function of error in the cross-validation, and upon a determination that the neural network is not overfit, deploying the trained neural network for production prediction operation;

determining a customer quality improvement rate (CQIR) for one or more customers, and a CQIR threshold, wherein the CQIR is a slope of a quality prediction line through at least two points determined by the processor based on customer data and an output of the trained neural network, the slope defining a rate at which a customer quality changes over time and the CQIR threshold is a rate of change of CQIR required to determine whether an individual customer meets desired potential customer quality trends, wherein the CQIR and CQIR threshold are based at least in part on use of said trained neural network; and

predicting future customer quality based at least in part on the CQIR and the trained neural network, wherein future customer quality is determined as a trend based on predicted quality evaluated for at least two points in time, and automatically rejecting customers whose CQIR does not satisfy the CQIR threshold, thereby filtering customers from a pool of potential customers based on the deployed trained neural network.

17 . The process of claim 16 , wherein the future customer quality trend is determined as a function of the slope of a quality prediction line.

18 . The process of claim 16 , wherein the future customer quality trend is determined as a function of CQIR.

19 . The process of claim 16 , wherein customer quality is determined as a function of input from a property management system based on image data.