Robust modeling

Details Robust modeling Robust modeling

1999-10-14

2003-02-18

Black, Thomas (Department: 2121)

Data processing: artificial intelligence

Machine learning

C706S025000, C706S014000, C706S015000, C706S023000

Reexamination Certificate

active

06523015

ABSTRACT:

FIELD OF THE INVENTION
The present invention relates generally to a learning machine that models a system. More specifically, a robust modeling system that determines an optimum complexity for a given criteria is disclosed. The robust model of a system strikes a compromise between accurately fitting outputs in a known data set and effectively predicting outputs for unknown data.
BACKGROUND OF THE INVENTION
A learning machine is a device that maps an unknown set of inputs (X
1
, X
2
, . . . X
n
) which may be referred to as an input vector to an output Y. Y may be a vector or Y may be a single value. Appropriate thresholds may be applied to Y so that the input data is classified by the output Y. When Y is a number, then the process of associating Y with an input vector is referred to as scoring and when Y is thresholded into classes, then the process of associating Y with an input vector is referred to as classification. A learning machine models the system that generates the output from the input using a mathematical model. The mathematical model is trained using a set of inputs and outputs generated by the system. Once the mathematical model is trained using the system generated data, the model may be used to predict future outputs based on given inputs.
A learning machine can be trained or using various techniques.
Statistical Learning Theory
by Vladmir Vapnik, published by John Wiley and Sons, ©1998, which is herein incorporated by reference for all purposes, and
Advances in Kernel Methods: Support Vector Learning
) published by MIT Press ©1999, which is herein incorporated by reference for all purposes describe how a linear model having a high dimensional feature space can be developed for a system that includes a large number of input parameters and an output.
One example of a system that may be modeled is electricity consumption by a household over time. The output of the system is the amount of electricity consumed by a household and the inputs may be a wide variety of data associated with empirical electricity consumption such as day of the week, month, average temperature, wind speed, household income, number of persons in the household, time of day, etc. It might be desirable to predict future electricity consumption by households given different inputs. A learning machine can be trained to predict electricity consumption for various inputs using a training data set that includes sets of input parameters (input vectors) and outputs associated with the input parameters. A model trained using available empirical data can then be used to predict future outputs from different inputs.
An important measure of the effectiveness of a trained model is its robustness. Robustness is a measure of how well the model performs on unknown data after training. As a more and more complex model is used to fit the training data set, the aggregate error produced by the model when applied to the entire training set can be lowered all the way to zero, if desired. However, as the complexity or capacity of the model increases, the error that is experienced on input data that is not included in the training set increases. That is because, as the model gets more and more complex, it becomes strongly customized to the training set. As it exactly models the vagaries of the data in the training set, the model tends to lose its ability to provide useful generalized results for data not included in the training set.
FIG. 1A
illustrates a model that is complex but is not robust. The output of the model is illustrated by trace
102
. Trace
102
passes very close to all of the data points shown, which are included in the training set. However, because of the complex nature of curve
102
, it is unlikely to successfully approximate the output Y for values of X that are not in the training set.
FIG. 1B
is a graph illustrating a model that is very robust, but does not provide as good a fit as the model shown in FIG.
1
A. Curve
104
does not pass as close to the data points in the training set shown as Curve
102
did in FIG.
1
A. However, Curve
104
is more robust because future data points shown as circles are closer to Curve
104
than to Curve
102
. In general, there is a tradeoff between providing a better and better fit for the points included in a training data set and the likelihood of a good fit for other data points not included in the training data set. The ability of the model to provide a good fit for data points not included in the training set is determined by the model's robustness. The question of how to determine an appropriately complex model so that the tradeoff between a good fit of the training set and robustness is the subject of considerable research.
For example, U.S. Pat. No. 5,684,929 (hereinafter the “'929 patent”) issued to Cortes and Jackel illustrates one approach to determining an appropriate complexity for a model used to predict the output of a system. Cortes and Jackel teach that, if data is provided in a training set used to train a model and a test data set used to test the model, then an approximation of the percentage error expected for a given level of complexity using a training set of infinite size can be accurately estimated. Based on such an estimate, Cortes and Jackel teach that combining such an estimate with other estimates obtained for different levels of capacity or complexity models can be used so that the error decreases asymptotically towards some minimum error E
m
. Cortes and Jackel then describe increasing the complexity of the modeling machine until the diminishing gains realized as the theoretical error for an infinite training set is asymptotically approached decrease below a threshold. The threshold may be adjusted to indicate when further decrease in error does not warrant increasing the complexity of the modeling function.
For very large training sets where the error on the test data set and the training data set both approximate the error on an infinite training set, this approach is useful. Generally, as complexity increases, the error decreases and it is reasonable to specify a minimum decrease in error below which it is not deemed worthwhile to further increase the complexity of the modeling function. However, the technique taught by Cortes and Jackel does not address the problem of the possible tradeoff in error for new data that results in error actually increasing as the modeling function complexity increases. By assuming that the training set is very large or perhaps infinite, if necessary, the '929 patent assumes that the error asymptotically reaches a minimum. That is not the case for finite data sets and therefore the phenomenon of reduced robustness with increased complexity should be addressed in practical systems with limited training data. What is needed is a way of varying the capacity or complexity of a modeling function and determining an optimum complexity for modeling a given system.
FIG. 2
is a graph illustrating how the error for a training data set and the error for data not included in the training set behave as the complexity of a model derived using the training data set increases. Curve
200
shows that as the complexity or capacity of the modeling function increases, the aggregate error calculated when comparing the output of the model to the output provided in the training data set for the same inputs decreases. In fact, the difference between the output of the model and the data provided in the training set can be reduced to zero if a sufficiently complex modeling function is used. Curve
202
illustrates the error determined by the difference between the output of the model and real output data obtained for inputs not included in the training set. As the complexity of the model increases, the error at first decreases until it reaches a minimum and then begins to increase. This result is caused by an overly complex model becoming excessively dependent on the vagaries of the training set. This phenomenon is referred to as over-training and results in a complex model that is a very good fi

Affiliated with

Also associated with

Rating

Robust modeling does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Robust modeling, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Robust modeling will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3180770