System and method for on-line training of a non-linear model for use in electronic commerce

Abstract

A system and method for on-line training of a non-linear model for use in electronic commerce. The non-linear model is trained with training sets from a stream of process data. The system detects availability of new training data, and constructs a training set from the corresponding input data. Over time, many training sets are presented to the non-linear model. When multiple presentations are needed to effectively train the non-linear model, a buffer of training sets is filled and updated as new training data becomes available. Once the buffer is full, a new training set bumps the oldest training set from the buffer. The training sets are presented one or more times each time a new training set is constructed. An historical database may be used to construct training sets for the non-linear model. The non-linear model may be trained retrospectively by searching the historical database and constructing training sets.

What is claimed is:  
     
         1 . A system adapted for predicting output data provided to a computer system used to control an electronic commerce system, the system comprising: 
 a processor;    a memory medium coupled to the processor, wherein the memory medium stores a non-linear model software program, wherein the non-linear model software program is coupled to retrieve training input data from a data source, and is coupled to retrieve input data from the data source in accordance with time specifications, and is operable to generate output data;    wherein the non-linear model software program comprises: 
 (a) specifications for said training input data, said input data, and said output data;  
 (b) coefficients for said non-linear model;  
 (c) program instructions for adjusting said coefficients in response to said training input data;  
 (d) program instructions for predicting said output data in accordance with said input data and said coefficients; and  
 (e) controlling the electronic commerce system using the predicted output data.  
   
     
     
         2 . The system of  claim 1 , wherein the electronic commerce system is an e-marketplace.  
     
     
         3 . The system of  claim 1 , further comprising: 
 a historical database, operable to provide one or more of said training input data and said input data, and operable to store a history of said output data with associated timestamps.    
     
     
         4 . The system of  claim 3 , wherein the non-linear model software program further comprises: 
 non-linear model specifications;    program instructions for entering said non-linear model specifications, wherein said program instructions implement:    a template presenting a partial non-linear model specification;    wherein said template is operable to receive user input specifying said non-linear model specifications wherein said user specified non-linear model specifications and said partial non-linear model specification specify the non-linear model.    
     
     
         5 . The system of  claim 4 , further comprising: 
 program instructions for sequencing, responsive to data specifications, and executable to retrieve data in accordance with said data specifications;    wherein said program instructions for sequencing are further executable to control execution of the non-linear model in accordance with said data specifications.    
     
     
         6 . The system of  claim 4 , further comprising: 
 program instructions for timing, responsive to data specifications, and executable to retrieve data in accordance with said data specifications;    wherein said program instructions for timing are further executable to: 
 detect new training input data;  
 determine said data specifications for said input data;  
 initiate training of the non-linear model; and  
 control the execution of said non-linear model, in accordance with time specifications.  
   
     
     
         7 . The system of  claim 6 , further comprising: 
 program instructions for feedback, wherein said program instructions for timing are further executable to control the execution of said program instructions for feedback.    
     
     
         8 . The system of  claim 6 , further comprising: 
 an input mechanism operable to sense a condition in the electronic commerce system;    an output mechanism operable to change a controllable state of the electronic commerce system;    said computer system operable to use data from said input mechanism as its input data for computing output data in accordance with said input data and in accordance with one or more parameters, and operable to send said output data to said output mechanism; and    wherein at least one of said non-linear model and said program instructions for feedback is operable to change said one or more parameters of said computer system.    
     
     
         9 . The system of  claim 6 , further comprising: 
 program instructions for an expert system, wherein said program instructions for timing are further executable to control execution of said program instructions for the expert system.    
     
     
         10 . The system of  claim 6 , further comprising: 
 program instructions for feedforward, wherein said program instructions for timing are further executable to control execution of said program instructions for feedforward.    
     
     
         11 . The system of  claim 6 , further comprising: 
 program instructions for statistical testing, wherein said program instructions for timing are further executable to control execution of said program instructions for statistical testing.    
     
     
         12 . The system of  claim 6 , further comprising: 
 program instructions for event processing, wherein said program instructions for timing are further executable to control execution of said program instructions for event processing.    
     
     
         13 . The system of  claim 1 , wherein the non-linear model software program is a support vector machine software program, wherein the support vector machine software program further comprises: 
 a support vector machine;    wherein the support vector machine further comprises support vector machine specifications;    wherein the support vector machine specifications further comprise specifications for a kernel function which operates as a basis finction for the support vector machine.    
     
     
         14 . The system of  claim 1 , wherein the non-linear model software program is a neural network software program, wherein the neural network software program further comprises: 
 a neural network;    wherein the neural network further comprises neural network specifications.    
     
     
         15 . The system of  claim 1 , wherein said predicting output data occurs substantially in real-time.  
     
     
         16 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control an electronic commerce system, the computer non-linear model method comprising: 
 (1) training a non-linear model using a first training set based on first data;    (2) training or retraining said non-linear model using a second training set based on second data, and using said first training set; and    (3) training or retraining said non-linear model using a third training set based on third data, and using said second training set, without using said first training set; and    (4) controlling the electronic commerce system using the predicted output data.    
     
     
         17 . The method of  claim 16 , wherein the electronic commerce system is an e-marketplace.  
     
     
         18 . The method of  claim 16 , wherein (1), (2), and/or (3) comprise retrieving said first training set, said second training set, or said third training set, respectively, from a database.  
     
     
         19 . The method of  claim 16 , wherein (1), (2), and/or (3) comprise retrieving said first training set, said second training set, or said third training set, respectively, from a historical database.  
     
     
         20 . The method of  claim 16 , wherein (1), (2), and/or (3) comprise constructing said first training set, said second training set, or said third training set, respectively.  
     
     
         21 . The method of  claim 20 , wherein said constructing operates substantially in real-time.  
     
     
         22 . The method of  claim 20 , wherein said constructing comprises using an associated timestamp(s) of said first data, said second data or said third data to indicate input data for constructing said first training set, said second training set, or said third training set, respectively.  
     
     
         23 . The method of  claim 16 , 
 wherein (1) is preceded by analyzing the process; and    wherein (1) further comprises using data representative of said analyzing as said first data.    
     
     
         24 . The method of  claim 16 , wherein the non-linear model is a support vector machine, wherein the support vector machine further comprises: 
 support vector machine specifications;    wherein the support vector machine specifications further comprise specifications for a kernel function which operates as a basis function for the support vector machine.    
     
     
         25 . The method of  claim 16 , wherein the non-linear model is a neural network, wherein the neural network further comprises: 
 neural network specifications.    
     
     
         26 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control an electronic commerce system, the computer non-linear model method comprising: 
 (1) detecting first data;    (2) training or retraining a non-linear model using a first training set based on said first data;    (3) detecting second data;    (4) training or retraining said non-linear model using a second training set based on said second data, and using said first training set;    (5) detecting third data;    (6) training or retraining said non-linear model using a third training set based on said third data, and using said second training set; and    (7) controlling the electronic commerce system using the predicted output data.    
     
     
         27 . The method of  claim 26 , wherein the electronic commerce system is an e-marketplace.  
     
     
         28 . The method of  claim 26 , further comprising retrieving said first training set, said second training set, and/or said third training set from a historical database.  
     
     
         29 . The method of  claim 26 , further comprising between steps (4) and (5) the step of discarding said first training set.  
     
     
         30 . The method of  claim 26 , further comprising after step (6) the step of discarding said second training set.  
     
     
         31 . The method of  claim 26 , wherein the non-linear model is a support vector machine, wherein the support vector machine further comprises: 
 support vector machine specifications;    wherein the support vector machine specifications further comprise specifications for a kernel function which operates as a basis function for the support vector machine.    
     
     
         32 . The method of  claim 26 , wherein the non-linear model is a neural network, wherein the neural network further comprises: 
 neural network specifications.    
     
     
         33 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control an electronic commerce system, the computer non-linear model method comprising: 
 (1) constructing a stack containing at least two training sets;    (2) training or retraining said non-linear model using said at least two training sets in said stack;    (3) constructing a new training set and replacing an oldest training set in said stack with said new training set; and    (4) repeating steps (2) and (3) at least once; and    (5) controlling the electronic commerce system using the predicted output data.    
     
     
         34 . The method of  claim 33 , wherein the electronic commerce system is an e-marketplace.  
     
     
         35 . The method of  claim 33 , wherein step (3) comprises: 
 monitoring substantially in real-time for the presence of new training input data; and    retrieving input data indicated by said new training input data to construct said new training set.    
     
     
         36 . The method of  claim 33 , wherein step (2) uses said at least two training sets of said stack one or more times.  
     
     
         37 . The method of  claim 33 , wherein the non-linear model is a support vector machine, wherein the support vector machine further comprises: 
 support vector machine specifications;    wherein the support vector machine specifications further comprise specifications for a kernel function which operates as a basis function for the support vector machine.    
     
     
         38 . The method of  claim 33 , wherein the non-linear model is a neural network, wherein the neural network further comprises: 
 neural network specifications.    
     
     
         39 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control an electronic commerce system, the computer non-linear model method comprising: 
 (1) operating the electronic commerce system and measuring the electronic commerce system to produce a first data, a second data, and a third data;    (2) training a non-linear model using a first training set based on said first data;    (3) training or retraining said non-linear model using a second training set based on said second data, and using said first training set; and    (4) training or retraining said non-linear model using a third training set based on said third data, and using said second training set; and    (5) controlling the electronic commerce system using the predicted output data.    
     
     
         40 . The method of  claim 39 , wherein the electronic commerce system is an e-marketplace.  
     
     
         41 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control an electronic commerce system, the computer non-linear model method comprising: 
 (1) training a non-linear model using a first training set based on first data;    (2) training or retraining said non-linear model using a second training set based on second data, and using said first training set;    (3) training or retraining said non-linear model using a third training set based on third data, and using said second training set;    (4) said non-linear model predicting a first output data using said first data; and    (5) changing a state of an output mechanism in accordance with said first output data; and    (6) controlling the electronic commerce system using the predicted output data.    
     
     
         42 . The method of  claim 41 , wherein the electronic commerce system is an e-marketplace.  
     
     
         43 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control an electronic commerce system, the computer non-linear model method comprising: 
 (1) detecting first data;    (2) training or retraining a non-linear model using a first training set based on said first data;    (3) detecting second data;    (4) training or retraining said non-linear model using a second training set based on said second data and by using said first training set;    (5) detecting third data;    (6) training or retaining said non-linear model using a third training set based on said third data, and using said second training set;    (7) said non-linear model predicting a first output data using said first data; and    (8) changing a state of an output mechanism in accordance with said first output data; and    (9) controlling the electronic commerce system using the predicted output data.    
     
     
         44 . The method of  claim 43 , wherein the electronic commerce system is an e-marketplace.  
     
     
         45 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control an electronic commerce system, the computer non-linear model method comprising: 
 (1) operating the electronic commerce system and measuring the electronic commerce system to produce a first data, a second data, and a third data;    (2) detecting said first data;    (3) training or retraining a non-linear model using a first training set based on said first data;    (4) detecting second data;    (5) training or retraining said non-linear model using a second training set based on said second data and using said first training set;    (6) detecting third data; and    (7) training or retraining said non-linear model using a third training set based on said third data, and using said second training set; and    (8) controlling the electronic commerce system using the predicted output data.    
     
     
         46 . The method of  claim 45 , wherein the electronic commerce system is an e-marketplace.  
     
     
         47 . A computer non-linear model system adapted for predicting output data provided to a computer system used to control an electronic commerce system, the computer non-linear model system comprising: 
 a non-linear model, connected to retrieve training input data from a data source, and connected to retrieve input data from the data source in accordance with time specifications, and connected to store output data to the data source, said non-linear model comprising: 
 (a) a memory for storing non-linear model specifications, said training input data, said input data, and said output data;  
 (b) the memory storing coefficients for said non-linear model;  
 (c) adjusting said coefficients in response to the training input data; and  
 (d) predicting said output data in accordance with said input data and said coefficients substantially in real-time; and  
 (e) controlling the electronic commerce system using the predicted output data.  
   
     
     
         48 . The method of  claim 47 , wherein the electronic commerce system is an e-marketplace.  
     
     
         49 . A modular computer non-linear model system adapted for predicting output data provided to a computer system used to control an electronic commerce system, the modular computer non-linear model system comprising: 
 (1) at least one module, comprising: 
 (i) at least one non-linear model, connected to retrieve training input data from a data source, and connected to retrieve input data from the data source in accordance with time specifications, and connected to store said output data to the data source, said non-linear model comprising: 
 (a) a memory for storing non-linear model specifications, said training input data, said input data, and said output data;  
 (b) the memory storing coefficients for said non-linear model;  
 (c) adjusting said coefficients in response to the training input data;  
 (d) predicting said output data in accordance with said input data and said coefficients substantially in real-time; and  
 
 (ii) modular timing and sequencing, responsive to said non-linear model specifications and said data specifications, and connected to retrieve data in accordance with said data specifications, said modular timing and sequencing comprising: 
 (a) detecting new training input data;  
 (b) determining said time specifications for said input data; and  
 (c) initiating training, said initiating connected to adjusting of said coefficients substantially in real-time; and  
 
   (2) controlling the electronic commerce system using the predicted output data.    
     
     
         50 . The method of  claim 49 , wherein the electronic commerce system is an e-marketplace.  
     
     
         51 . A computer non-linear model system adapted for predicting output data provided to a computer system used to control an electronic commerce system, the computer non-linear model system comprising: 
 (1) a memory for storing one or more training sets;    (2) a non-linear model, connected to retrieve said training sets from said memory and connected to store output data, comprising: 
 (a) the memory storing non-linear model specifications, said training sets, and said output data;  
 (b) the memory storing coefficients for said non-linear model;  
 (c) adjusting said coefficients in accordance with said training sets; and  
 (d) predicting said output data in accordance with said training sets and said coefficients substantially in real-time;  
   (3) constructing said non-linear model in response to said training sets, said constructing comprising: 
 (a) detecting new training input data;  
 (b) determining time specifications for said new training input data;  
 (c) retrieving said new training input data, and input data in accordance with said time specifications;  
 (d) removing an existing training set from said memory;  
 (e) storing said retrieved new training input data and said input data as a new training set in said memory; and  
 (f) initiating training, said initiating connected to adjusting of said coefficients substantially in real-time; and  
   (4) controlling the electronic commerce system using the predicted output data.    
     
     
         52 . The method of  claim 51 , wherein the electronic commerce system is an e-marketplace.  
     
     
         53 . A system adapted for predicting output data provided to a computer system used to control a financial process, the system comprising: 
 a processor;    a memory medium coupled to the processor, wherein the memory medium stores a non-linear model software program, wherein the non-linear model software program is coupled to retrieve training input data from a data source, and is coupled to retrieve input data from the data source in accordance with time specifications, and is operable to generate output data;    wherein the non-linear model software program comprises: 
 (a) specifications for said training input data, said input data, and said output data;  
 (b) coefficients for said non-linear model;  
 (c) program instructions for adjusting said coefficients in response to the training input data; and  
 (d) program instructions for predicting said output data in accordance with said input data and said coefficients; and  
 (e) controlling the financial process using the predicted output data.  
   
     
     
         54 . The system of  claim 53 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.  
     
     
         55 . The system of  claim 53 , further comprising: 
 a historical database, operable to provide one or more of said training input data and said input data, and operable to store a history of said output data with associated timestamps.    
     
     
         56 . The system of  claim 55 , wherein the non-linear model software program further comprises: 
 non-linear model specifications;    program instructions for entering said non-linear model specifications, wherein said program instructions implement:    a template presenting a partial non-linear model specification;    wherein said template is operable to receive user input specifying said non-linear model specifications wherein said user specified non-linear model specifications and said partial non-linear model specification specify the non-linear model.    
     
     
         57 . The system of  claim 56 , further comprising: 
 program instructions for sequencing, responsive to data specifications, and executable to retrieve data in accordance with said data specifications;    wherein said program instructions for sequencing are further executable to control execution of the non-linear model in accordance with said data specifications.    
     
     
         58 . The system of  claim 56 , further comprising: 
 program instructions for timing, responsive to data specifications, and executable to retrieve data in accordance with said data specifications;    wherein said program instructions for timing are further executable to: 
 detect new training input data;  
 determine said data specifications for said input data;  
 initiate training of the non-linear model; and  
 control the execution of said non-linear model, in accordance with time specifications.  
   
     
     
         59 . The system of  claim 58 , further comprising: 
 program instructions for feedback, wherein said program instructions for timing are further executable to control the execution of said program instructions for feedback.    
     
     
         60 . The system of  claim 58 , further comprising: 
 an input mechanism operable to sense a condition in the financial process;    an output mechanism operable to change a controllable state of the financial process;    said computer system operable to use data from said input mechanism as its input data for computing output data in accordance with said input data and in accordance with one or more parameters, and operable to send said output data to said output mechanism; and    wherein at least one of said non-linear model and said program instructions for feedback is operable to change said one or more parameters of said computer system.    
     
     
         61 . The system of  claim 58 , further comprising: 
 program instructions for an expert system, wherein said program instructions for timing are further executable to control execution of said program instructions for the expert system.    
     
     
         62 . The system of  claim 58 , further comprising: 
 program instructions for feedforward, wherein said program instructions for timing are further executable to control execution of said program instructions for feedforward.    
     
     
         63 . The system of  claim 58 , further comprising: 
 program instructions for statistical testing, wherein said program instructions for timing are further executable to control execution of said program instructions for statistical testing.    
     
     
         64 . The system of  claim 58 , further comprising: 
 program instructions for event processing, wherein said program instructions for timing are further executable to control execution of said program instructions for event processing.    
     
     
         65 . The system of  claim 53 , wherein the non-linear model software program is a support vector machine software program, wherein the support vector machine software program further comprises: 
 a support vector machine;    wherein the support vector machine further comprises support vector machine specifications;    wherein the support vector machine specifications further comprise specifications for a kernel function which operates as a basis function for the support vector machine.    
     
     
         66 . The system of  claim 53 , wherein the non-linear model software program is a neural network software program, wherein the neural network software program further comprises: 
 a neural network;    wherein the neural network further comprises neural network specifications.    
     
     
         67 . The system of  claim 53 , wherein said predicting output data occurs substantially in real-time.  
     
     
         68 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control a financial process, the computer non-linear model method comprising: 
 (1) training a non-linear model using a first training set based on first data;    (2) training or retraining said non-linear model using a second training set based on second data, and using said first training set; and    (3) training or retraining said non-linear model using a third training set based on third data, and using said second training set, without using said first training set; and    (4) controlling the financial process using the predicted output data.    
     
     
         69 . The method of  claim 68 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.  
     
     
         70 . The method of  claim 68 , wherein (1), (2), and/or (3) comprise retrieving said first training set, said second training set, or said third training set, respectively, from a database.  
     
     
         71 . The method of  claim 68 , wherein (1), (2), and/or (3) comprise retrieving said first training set, said second training set, or said third training set, respectively, from a historical database.  
     
     
         72 . The method of  claim 68 , wherein (1), (2), and/or (3) comprise constructing said first training set, said second training set, or said third training set, respectively.  
     
     
         73 . The method of  claim 72 , wherein said constructing operates substantially in real-time.  
     
     
         74 . The method of  claim 72 , wherein s aid constructing comprises using an associated timestamp(s) of said first data, said second data or said third da ta to indicate input data for constructing said first training set, said second training set, or said third training set, respectively.  
     
     
         75 . The method of  claim 68 , 
 wherein (1) is preceded by analyzing the financial process; and    wherein (1) further comprises using data representative of said analyzing as said first data.    
     
     
         76 . The method of  claim 68 , wherein the non-linear model is a support vector machine, wherein the support vector machine further comprises: 
 support vector machine specifications;    wherein the support vector machine specifications further comprise specifications for a kernel function which operates as a basis function for the support vector machine.    
     
     
         77 . The method of  claim 68 , wherein the non-linear model is a neural network, wherein the neural network further comprises: 
 neural network specifications.    
     
     
         78 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control a financial process, the computer non-linear model method comprising: 
 (1) detecting first data;    (2) training or retraining a non-linear model using a first training set based on said first data;    (3) detecting second data;    (4) training or retraining said non-linear model using a second training set based on said second data, and using said first training set;    (5) detecting third data;    (6) training or retraining said non-linear model using a third training set based on said third data, and using said second training set; and    (7) controlling the financial process using the predicted output data.    
     
     
         79 . The method of  claim 78 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.  
     
     
         80 . The method of  claim 78 , further comprising retrieving said first training set, said second training set, and/or said third training set from a historical database.  
     
     
         81 . The method of  claim 78 , further comprising between steps (4) and (5) the step of discarding said first training set.  
     
     
         82 . The method of  claim 78 , further comprising after step (6) the step of discarding said second training set.  
     
     
         83 . The method of  claim 78 , wherein the non-linear model is a support vector machine, wherein the support vector machine further comprises: 
 support vector machine specifications;    wherein the support vector machine specifications further comprise specifications for a kernel function which operates as a basis function for the support vector machine.    
     
     
         84 . The method of  claim 78 , wherein the non-linear model is a neural network, wherein the neural network further comprises: 
 neural network specifications.    
     
     
         85 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control a financial process, the computer non-linear model method comprising: 
 (1) constructing a stack containing at least two training sets;    (2) training or retraining said non-linear model using said at least two training sets in said stack;    (3) constructing a new training set and replacing an oldest training set in said stack with said new training set; and    (4) repeating steps (2) and (3) at least once; and    (5) controlling the financial process using the predicted output data.    
     
     
         86 . The method of  claim 85 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.  
     
     
         87 . The method of  claim 85 , wherein step (3) comprises: 
 monitoring substantially in real-time for the presence of new training input data; and    retrieving input data indicated by said new training input data to construct said new training set.    
     
     
         88 . The method of  claim 85 , wherein step (2) uses said at least two training sets of said stack one or more times.  
     
     
         89 . The method of  claim 85 , wherein the non-linear model is a support vector machine, wherein the support vector machine further comprises: 
 support vector machine specifications;    wherein the support vector machine specifications further comprise specifications for a kernel function which operates as a basis function for the support vector machine.    
     
     
         90 . The method of  claim 85 , wherein the non-linear model is a neural network, wherein the neural network further comprises: 
 neural network specifications.    
     
     
         91 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control a financial process, the computer non-linear model method comprising: 
 (1) operating the financial process and measuring the financial process to produce a first data, a second data, and a third data;    (2) training a non-linear model using a first training set based on said first data;    (3) training or retraining said non-linear model using a second training set based on said second data, and using said first training set; and    (4) training or retraining said non-linear model using a third training set based on said third data, and using said second training set; and    (5) controlling the financial process using the predicted output data.    
     
     
         92 . The method of  claim 91 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.  
     
     
         93 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control a financial process, the computer non-linear model method comprising: 
 (1) training a non-linear model using a first training set based on first data;    (2) training or retraining said non-linear model using a second training set based on second data, and using said first training set;    (3) training or retraining said non-linear model using a third training set based on third data, and using said second training set;    (4) said non-linear model predicting a first output data using said first data; and    (5) changing a state of an output mechanism in accordance with said first output data; and    (6) controlling the financial process using the predicted output data.    
     
     
         94 . The method of  claim 93 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.  
     
     
         95 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control a financial process, the computer non-linear model method comprising: 
 (1) detecting first data;    (2) training or retraining a non-linear model using a first training set based on said first data;    (3) detecting second data;    (4) training or retraining said non-linear model using a second training set based on said second data and by using said first training set;    (5) detecting third data;    (6) training or retaining said non-linear model using a third training set based on said third data, and using said second training set;    (7) said non-linear model predicting a first output data using said first data; and    (8) changing a state of an output mechanism in accordance with said first output data; and    (9) controlling the financial process using the predicted output data.    
     
     
         96 . The method of  claim 95 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.  
     
     
         97 . A computer non-linear model method adapted for predicting output data provided to a computer system used to control a financial process, the computer non-linear model method comprising: 
 (1) operating the financial process and measuring the financial process to produce a first data, a second data, and a third data;    (2) detecting said first data;    (3) training or retraining a non-linear model using a first training set based on said first data;    (4) detecting second data;    (5) training or retraining said non-linear model using a second training set based on said second data and using said first training set;    (6) detecting third data; and    (7) training or retraining said non-linear model using a third training set based on said third data, and using said second training set; and    (8) controlling the financial process using the predicted output data.    
     
     
         98 . The method of  claim 97 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.  
     
     
         99 . A computer non-linear model system adapted for predicting output data provided to a computer system used to control a financial process, the computer non-linear model system comprising: 
 a non-linear model, connected to retrieve training input data from a data source, and connected to retrieve input data from the data source in accordance with time specifications, and connected to store output data to the data source, said non-linear model comprising: 
 (a) a memory for storing non-linear model specifications, said training input data, said input data, and said output data;  
 (b) the memory storing coefficients for said non-linear model;  
 (c) adjusting said coefficients in response to the training input data; and  
 (d) predicting said output data in accordance with said input data and said coefficients substantially in real-time; and  
   controlling the financial process using the predicted output data.    
     
     
         100 . The method of  claim 99 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.  
     
     
         101 . A modular computer non-linear model system adapted for predicting output data provided to a computer system used to control a financial process, the modular computer non-linear model system comprising: 
 (1) at least one module, comprising: 
 (i) at least one non-linear model, connected to retrieve training input data from a data source, and connected to retrieve input data from the data source in accordance with time specifications, and connected to store said output data to the data source, said non-linear model comprising: 
 (a) a memory for storing non-linear model specifications, said training input data, said input data, and said output data;  
 (b) the memory storing coefficients for said non-linear model;  
 (c) adjusting said coefficients in response to the training input data;  
 (d) predicting said output data in accordance with said input data and said coefficients substantially in real-time; and  
 
 (ii) modular timing and sequencing, responsive to said non-linear model specifications and said data specifications, and connected to retrieve data in accordance with said data specifications, said modular timing and sequencing comprising: 
 (a) detecting new training input data;  
 (b) determining said time specifications for said input data; and  
 (c) initiating training, said initiating connected to adjusting of said coefficients substantially in real-time; and  
 
   (2) controlling the financial process using the predicted output data.    
     
     
         102 . The system of  claim 101 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.  
     
     
         103 . A computer non-linear model system adapted for predicting output data provided to a computer system used to control a financial process, the computer non-linear model system comprising: 
 (1) a memory for storing one or more training sets;    (2) a non-linear model, connected to retrieve said training sets from said memory and connected to store output data, comprising: 
 (a) the memory storing non-linear model specifications, said training sets, and said output data;  
 (b) the memory storing coefficients for said non-linear model;  
 (c) adjusting said coefficients in accordance with said training sets; and  
 (d) predicting said output data in accordance with said training sets and said coefficients substantially in real-time;  
   (3) constructing said non-linear model in response to said training sets, said constructing comprising: 
 (a) detecting new training input data;  
 (b) determining time specifications for said new training input data;  
 (c) retrieving said new training input data, and input data in accordance with said time specifications;  
 (d) removing an existing training set from said memory;  
 (e) storing said retrieved new training input data and said input data as a new training set in said memory; and  
 (f) initiating training, said initiating connected to adjusting of said coefficients substantially in real-time; and  
   (4) controlling the financial process using the predicted output data.    
     
     
         104 . The system of  claim 103 , wherein the financial process is one of: a financial analysis process, a portfolio management process, a bond analysis process, a stock analysis process.