Profiles are conceptual topics that are created automatically for an individual, based on their usage of the system. Using information such as their searches, or documents viewed and created, an individual typically has multiple profiles that evolve over time. Profiles are then used for a number of operations such as automatic delivery of relevant content.
IDOL Server can automatically create interest and expertise profiles for your users, based on their document viewing history.
As IDOL Server builds the profile, it creates agents. Each agent defines a topic that the user is interested in. You can in turn use the agents to return documents that match the user’s interests and expertise. You can also manually edit the agents to improve the results, if required.
Profiling provides an automatic process for finding out what a user is interested in, and it allows you (or the user) to return documents tailored to user preferences.
You can create multiple profiles for users. For example, you might want to set up a personal profile and a professional profile, to keep their business and personal preferences separate.
To use multiple profiles, you use named areas. For this example, you have a professional and personal named area. When the user views documents related to their work projects, it updates the profile in the professional named area. When they view documents related to their hobbies, it updates the profile in the personal named area.
Profiles make it possible to return documents that are immediately relevant to your users, by using the preferences that they have implicitly expressed. It also means that users receive more specific information, according to their preferences.
You can use profiles in a number of ways. For example, you might want to:
make product recommendations to a Web site user, based on their browsing history. For example, you can use it to promote products that are similar to things the user has previously bought.
boost search results according to a user’s previous searches.
make article recommendations on a Web site, according to the articles that they have previously read.
alert a user to the release of new products that are similar to products that they have previously bought.
|
|