Automated learning of models for domain theories

1. A computer implemented method for determining a model, the method comprising:

(a) identifying, by one or more computing devices, a schema that defines a possible causal element of a particular type of behavior;

(b) determining, by the one or more computing devices, one or more concepts provided in the schema, and one or more sub-concepts for each concept, each concept and each sub-concept being associated with uncalibrated weights and a functional form which represent a logical relationship between them;

(c) determining, by the one or more computing devices, multiple models from the one or more concepts and the one or more sub-concepts;

(d) calibrating, by the one or more computing devices, the multiple models using representative data collected from a real-world source, wherein the calibration utilizes a limited-memory Broyden-Fletcher-Goldfarb-Shanno algorithm to iteratively evaluate a series of possible weight values for a given combination of sub-concepts;

(e) updating the uncalibrated weights with weights from the possible weight values to determine an optimal model amongst a plurality of calibrated models;

(f) determining, by the one or more computing devices, the optimal model amongst the plurality of calibrated models based on a model score that includes a determination of a model fit and a compliance with an expected attribution associated with an existing knowledge of the schema, wherein the model fit is calculated based on real-time data from the real-world source, and wherein determining the optimal model includes adaptively learning the optimal model by repeating (c) and (d) to converge on the optimal model; and

(g) transmitting, by the one or more computing devices, the optimal model to a user interface to thereby support measurement of marketing effectiveness in response to one or more stimuli in a marketplace.

2. The method of claim 1 , wherein identifying the schema includes identifying an independent variable.

3. The method of claim 1 , wherein the schema is associated with a termination threshold.

4. The method of claim 3 , further comprising:

after calibrating the model, making a determination as to whether the model satisfies the termination threshold.

5. The method of claim 1 , wherein determining the optimal model includes selecting the optimal model from the plurality of calibrated models.

6. The method of claim 1 , wherein the model score is based on the model fit to a real-world result.

7. The method of claim 1 , wherein the one or more concepts are based on one or more rules that are specific to a domain threshold.

8. The method of claim 1 , wherein a user identifies the schema by mapping real-world data to each of a dependent and independent variables.

9. The method of claim 1 , wherein a user is separated from the representative data, such that the representative data is inaccessible to the user.

10. The method of claim 1 , wherein the determining the optimal model includes using a genetic algorithm.

11. The method of claim 10 , wherein (d) is performed using a statistical process.

12. The method of claim 1 , further comprising determining an alternative model by forming a cross-over combination of at least two of the plurality of calibrated models, the forming including exchanging at least one feature of the at least two of the plurality of calibrated models to form the alternative model.

13. The method of claim 1 , further comprising determining an alternative model by using a mutation process, wherein one or more features of the at least one of the plurality of calibrated models are randomly altered with a probability equal to a mutation rate.

14. The method of claim 1 , further comprising adjusting a marketing activity based on the optimal model.