Exponential priors for maximum entropy models

• US 7,184,929 B2
• Filed: 01/28/2004
• Issued: 02/27/2007
• Est. Priority Date: 01/28/2004
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Goodman, Joshua T.
• Primary Examiner(s): Tsai; Carol S. W.
• Assistant Examiner(s): Huynh; Phuong
• Application Number:
  US10/766,348
• Publication Number:
  US 20050165580A1
• Time in Patent Office: 1,126 Days
• Field of Search: 702180-182, 702/184, 702/179, 706/21, 703/2, 709/206, 709/207, 709/224
• US Class Current: 702/181
• CPC Class Codes: G06F 18/21 Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 20/10 using kernel methods, e.g. support vector machines [SVM]

Abstract

The subject invention provides for systems and methods that facilitate optimizing one or mores sets of training data by utilizing an Exponential distribution as the prior on one or more parameters in connection with a maximum entropy (maxent) model to mitigate overfitting. Maxent is also known as logistic regression. More specifically, the systems and methods can facilitate optimizing probabilities that are assigned to the training data for later use in machine learning processes, for example. In practice, training data can be assigned their respective weights and then a probability distribution can be assigned to those weights.

Claims

1. A computer implemented system that facilitates maximizing probabilities to effectuate spain filtering comprising:

a data input component that provides one or more types of data for analysis; and

an analysis component that analyzes at least a subset of the one or more types of data to compute maximized probabilities to identify at least one spain email message by employing an iterative scaling function, a plurality of Exponential priors that correspond to a plurality of different features respectively, and at least one of;

a LaPlacian prior and a non-Gaussian.

2. The system of claim 1, the iterative scaling function comprises generalized iterative scaling.

3. The system of claim 1, the iterative scaling function comprises improved iterative scaling.

4. The system of claim 1, the iterative scaling function comprises sequential generalized iterative scaling.

5. The system of claim 1, the plurality of Exponential priors depends on counts of the features.

6. The system of claim of 1, the plurality of Exponential priors depends in part upon a usefulness of a feature.

7. The system of claim 5, the counts are based in part upon a Good-Turing estimate.

8. The system of claim 1, the analysis component comprising:

a maximization component that provides instructions for computing a maximum value;

a model component operatively coupled to the maximization component that receives data from at least the maximization component and at least an Exponential prior component; and

a probability processing component that employs information collected by the model component to compute one or more values.

9. A computer-readable medium having stored thereon the following computer executable components to identify spain email messages, comprising:

a data input component that provides one or more types of data for analysis; and

an analysis component that analyzes at least a subset of the one or more types of data to compute maximized probabilities by employing at least an iterative scaling function, a plurality of Exponential priors that correspond to a plurality of different features respectively, and at least one LaPlacian prior and a non-Gaussian distribution, the maximized probabilities employed to ascertain that at least one of the one or more types of data includes a spain email message.