Content Sampling

Introduction

The Classification Intelligence feature of Records365 allows you to use machine learning techniques to make decisions based on the content created within your organization. Due to policies and or regulations you may be restricted as to where your content can be stored and for how long.

Records365 leverages the content under management to create highly accurate machine learning models. The content sampling feature allows you to reduce the size of the content footprint managed within Records365 and still provide a high probability of record category prediction.

How it works

When the Content Sampling feature is enabled, Records365 captures the minimum amount of content for Classification Intelligence purposes.  Records used for Content Sampling will have a ‘Retained’ state.

Content will be retained for the following reasons;

• The record was categorized by records classification and is within the latest 500 records for that record category. This includes newer versions of records that were previously cleared.
• The record is awaiting Classification Intelligence suggestion acceptance or rejection.
• The record has had a Classification Intelligence suggestion rejected.
• The record is part of the current Classification Intelligence training set.
• The record has been ingested since the last scheduled run of removing content.
• Less than 7 days since the content was created in Records365. 7 days is to allow Classification Intelligence to provide suggestions using all content for records with multiple associated binaries.

On a daily basis Records365 will review the content it has under management and remove content which;

• No longer meets the above ‘Retained’ criteria.
• Has been assigned a record category that is excluded from use by Classification Intelligence, refer to the Available for Classification Intelligence field on a record category. More information can be found on the File Plan page.

*Please note, those items where the suggestion was rejected will be included in the Classification Intelligence training set.

Those removed will then transition to a ‘Cleared’ state.

State information is available for searching using the Advanced Search page. For more information on this please see Record Content.