In one of our latest blogs, highlighting the “Top Innovations of 2020 for Oracle Content Management”, we talk briefly about the benefits of enhanced AI-enabled content. In a previous blog on Smart Content, we described how auto-tagging of images, text, Smart Search and Smart Authoring help with regards to improved content discoverability. In this follow up, we are going to take a deeper dive into this idea, specifically around Smart Categorization.
A pre-requisite to content discovery in any content management strategy is tagging and categorizing. Tagging and categorization used to be a painstakingly manual process, requiring a lot of time for initial set up, as well as for maintaining the accuracy as content evolves. In this blog we will focus on Smart categorization: How it works and How to use.
The goal of Smart Categorization is to help customers easily categorize content into the customer defined categories (taxonomy). Enterprises can have multiple taxonomies in a repository mirroring their business needs. When new content is added, or new taxonomies are created the content needs to be classified. This process of manual categorization is an expensive and error-prone process, especially as the amount of content and taxonomies grow. Our modelling shows that enterprises need to spend hundreds of hours to properly categorize content. Smart Categorization helps to alleviate this problem by providing category suggestions using AI-based recognition on existing categories and content in the repository.
This classification of content not only helps with organization and discoverability for authors but also helps deliver relevant content to consumers through personalization.
As customers categorize similar content to the same category of the taxonomy, patterns start to become identifiable. Given this state, the content residing inside an already classified taxonomy can be analyzed by an Artificial Intelligence (AI) algorithm to determine where a newly created/edited content item may belong. This AI algorithm continuously learns from the past data and generates category suggestions with the aim of reducing effort for customers.
In addition, the algorithm learns from suggestions that are accepted and suggestions that are rejected by the user to make more suitable suggestions.
Smart categorization can be used to categorize:
1. A single content item asset one at a time or
2. Categorize several content items at a time (Bulk categorization suggestions)
For single content item categorization:
1. Edit the content item and toggle the sidebar open
2. In the sidebar, choose “Categories” section
3. On the bottom bar, “View suggestions” is available. Clicking on it generates Classification suggestions for this Content item.
Fig 1: Single Item Classification
For help with bulk categorization:
1. Click Categories or open the More menu
2. Click View Category Suggestions under Categories or select Category Suggestions from the More menu. You can enter a category in the search field to quickly find the one you want, and sort the Categories with Suggestions listing by name, number of suggested content items per category, or by category path.
Fig 2: Bulk categorization Suggestions
Currently, we generate category suggestions only for content items. In upcoming releases, we will be providing category suggestions to additional document types like pdf, txt etc.