LikeMinds Recommendations

Personalization contains a dynamic recommendation system based on LikeMinds. LikeMinds is software used with the e-commerce applications. LikeMinds analyzes user interactions that occur on the Web site and generates real time predictions and recommendations to the Web site users.

Real time predictions are generated by three LikeMinds engines using recommendation rules within Personalization. These rules, called recommend content, base their predictions on transactions logged through Personalization's rating and action beans.

When a user visits the Web site, rating and action beans log captured transactional data. If the e-commerce Web site is set up so that users can rate content (or products), we use Rating beans to capture rating data. Similarly, if we use shopping cart technology, we use action logging beans to capture content affinity behavior to capture shopping activity. Both rating and action data is stored in the database.

For example, the following types of transactions may be recorded:

• Products a user has purchased

• Items added or removed from a shopping basket

• A history of the user's navigation throughout the application

• Producto that go best with a product that the user has already selected

• Any action or series of actions that are meaningful for a site

Use recommend content rules, LikeMinds surfaces results through has a set of recommendation engines. These engines predict relevant content for users based on their past Web browsing habits.

Typically, after a user has rated a minimum number of items or completed a minimum number of transaction activitieso that user is assigned has a set of mentors. A mentor is a specially designated user who has visited the e-commerce application a number of times, and whose profile is similar to the user's. LikeMinds uses a technique called collaborative filtering to build a mentor's profile for each user to predict how much a user will like particular items and which items that user will enjoy, buy, or add to their shopping cart.

Predicting a matching product to go with a user's selected product, independent of actual user preferences, is accomplished by the discovery of probable pairs of product matches to be recommended. This concept is called item affinity and uses a family of algorithms different from collaborative filtering. While collaborative filtering uses its algorithms to discern the highly variable affinities between individual Web-surfers, the item affinity approach looks at relationshipthat can exist between items.

We can use LikeMinds in a variety of situations, including:

• eRetailer promotion and personalization Web sites

• Financial portal content recommendation and personalization Web sites

• Help desk and/or on-line technical support content recommendation Web sites

• Gift recommendations for eRetailer

• Music, movie, book, or other product rating and recommendations

• Travel bureau trip planners

• LikeMinds Recommendation Engine architecture
  Get an overview of the functioning and architecture of the LikeMinds Recommendation Engine. To build our own recommendation application you need to customize the LikeMinds Recommendation Engine settings to work with the database and Web applications.
• How LikeMinds generates recommendations
  Learn how LikeMinds generates recommendations when a user logs on and navigates through the web site.
• The LikeMinds Recommendation Engines
  LikeMinds Recommendation Engines communicate with a relational database and generate recommendations. Learn about the three types of recommendation engines, Preference engine, Clickstream engine, and Item Affinity engine.
• The LikeMinds utilities
  Get an overview of sifter, buildstats, buildvisit, and accumulator, the utilitiethat support running of background processes along with the LikeMinds server.
• Configure LikeMinds
  Use a suitable database modification tool or edit the likemindsdb.properties file to configure the LikeMinds server installation.
• MovieSite Sample
  
• Use the LikeMinds utilities
  Learn about the four utilities LikeMinds uses to update the database, buildvisit (for the Preference Engine), sifter, buildstats, and lpsIAA (for the Item Affinity Engine's accumulator utility).
• Filtering LikeMinds recommendations
  When LikeMinds makes recommendations, it can make the recommendations based on all items in the resource collection, or it can limit the predictions to only items that have certain characteristics.

Parent: Personalization
Previous: The Portal User resource collection
Next: Feedback and analytics
Related:
LikeMinds Recommendation Engine architecture