The Generic Context Model of the system is consisted of all the context variables currently available in the system. CARS developers can use this generic model to choose context variables that are of interest to them and use them in their own application context models (application contexts). You may use the Options menu below to display application contexts or create an application context.

An Application Context is a context model for a particular application. It is built by a CARS developer in order to model the context for his/her application.

A Context instance is basically a "screenshot" of the context during an event of interaction between the user and the item that are involved in the recommendation process. For example, for a movie recommender a context instance is the set of context variables that constitute the context during the event of a particular user (user='Tom') watching a particular movie (item='Rambo') at a particular time (lets say the first time that Tom watches Rambo). Such a context instance may have title: 'Tom-Rambo C1' (C1 results from 'Context1') and may be consisted of the context variables: the time of day, the day of the week, the IMDB ratings of the movie watched, whom did the user watch the movie with, etc. In a similar way, the context instance around the second time Tom watches Rambo will have title 'Tom-Rambo C2' and will again be consisted of a number of context variables.

In addition, a context instance is a "screenshot" of the context for a single recommendation (of a particular recommender system). Therefore, a context instance can only exist within a particular applicatin context.

Therefore, a context instance is the set of context variables with their corresponding values that constitute the context for a single recommendation.

A context variable on the other hand may participate in a number of context instances, each of which is valid when the context variable is enabled. For example, the context variable "time: morning" can participate in many context instances, all of which refer to morning time.

Based on the above, the system’s context instances define all valid contextual information around a fact or event (in the example above around a particular user watching a particular movie at a certain time). Of course, context instances should be automatically created in some way (context sensing and retrieval) during the occurrence of the facts/events and stored in the system; e.g. around the event of a user watching a movie the system should be automatically aware of: the time of day, the day of the week, the IMDB ratings of the movie watched, whom did the user watch the movie with, and many more context variables that play a role in the process. This is a very difficult process and in some way impossible to achieve (how is the system going to know whom did the user watch the movie with, unless the user states it?). In this modelling system and learning tool we do not provide such functionality; however, we do provide you with the possibility as a CARS developer to create your own context instances for modelling purposes, in order to closely observe and study whether your application context is able to "catch" any context instance that may occur. We provide the functionality of creating a new context instance and displaying the context instances stored in the system.

A Context Dimension is a context variable with all of its values. Context Dimensions represent the additional dimensions that are being used in the recommendation process besides the user and the item. In this work we use the following definition of the context dimension: each individual context variable with a unique name and with all of its values can be perceived as a context dimension. Therefore a context dimension's name is a unique context variable name. For more information in the literature on context dimensions please refer to: "Adomavicius G., and Tuzhilin, A. 2008. Context-Aware Recommender Systems. Handbook on Recommender Systems, Springer, 2010, pages 10-11