Download VML profile feature set

book

Article ID: 159687

calendar_today

Updated On:

Products

Data Loss Prevention Endpoint Prevent Data Loss Prevention Network Prevent for Email Data Loss Prevention Enforce Data Loss Prevention Network Discover Data Loss Prevention Network Prevent for Web Data Loss Prevention Network Protect Data Loss Prevention for Tablets Data Loss Prevention Endpoint Discover

Issue/Introduction

Vector Machine Learning (VML) is detection technology available with Symantec Data Loss Prevention version 11.1 and later. For more information refer to the Symantec Data Loss Prevention Administrtion Guide and the Vector Machine Learning Best Practices Guide

To debug a VML profile, you can download and review its feature set. Features are keywords extracted from the training set documents. The system normalizes the keywords to lower case, removes white space and punctuation, classifies and plots the frequency of the features to generate the model.

Resolution

To download a VML profile model for debug purposes:

1. In Enforce select the VML profile you want to debug from the screen Manage > Data Profiles > Vector Machine Learning.

2. Look in the URL for the "mldProfileId=X" value.

3. Download the feature-set from Enforce using the following URL:

https://<EnforceServer>/ProtectManager/DownloadMachineLearningFeatures.do?mldProfileId=<ProfileId>&type=approvedResult

a) Replace <EnforceServer> with the Enforce server name.
b) Replace <ProfileId> with the VML profile ID.

For example:

https://localhost/ProtectManager/DownloadMachineLearningFeatures.do?mldProfileId=1&type=approvedResult

Where "mldProfileId=1" is the VML profile ID.

4. Download the file or open it in the browser. The feature-set file is named "features.txt."

Note: If you use IE 7 or 8 to download or view the model feature-set, the first time you attempt to do so it fails. Subsequent attempts should succeed. If not, use Firefox.

5. The file lists the features extracted during the creation of the VML profile's model. The numbers beside each feature indicate the following:

1  = positive feature
-1 = negative feature
0  = common feature