Abstract: A disclosed apparatus and method provide recommendations for products which a customer may need to order and/or items which the customer may have forgotten to order. An exemplary recommendation unit can communicate with a product information database, a product similarity database, a customer purchase history database and a customer order database. The recommendation unit includes a determining unit to store a determined similarity of products in the product similarity database, a training unit to train a model based on information about products previously purchased by each customer and information about at least one determined similarity as stored in the product similarity database, a similarity unit to generate at least one product similar to at least one product to be purchased by a customer based on information about a current order of a customer, and a calculating unit to calculate a probability that a customer forgot to add a similar product.