System and method for providing natural language recommendations

1. A stand-alone device that provides recommendations, the device comprising:

an input component that receives a request from a user;

a storage component, in communication with the input component, that stores an item attribute value for an item representing a preference of the user;

a processor, in communication with the storage component, that executes a recommendation engine to identify a recommendation based on the item attribute value, wherein the recommendation engine determines the item from parsing a natural language expression of a person other than the user; and

an output component, in communication with the processor, that provides a recommendation to the user.

2. The device of claim 1 wherein the storage component also stores a value of an environmental parameter in association with the item attribute value and the recommendation is further based on the value of the environmental parameter and the current environment.

3. The device of claim 2 wherein the environmental parameter is a type of activity.

4. The device of claim 1 wherein the storage component stores a user database, and the processor derives values of a user attribute vector by classifying the user based on attributes of other users in the user database such that the recommendation is further based on the user attribute vector.

5. The device of claim 1 further comprising a network interface for accessing a server, wherein the device syncs the user's preferences with the server when access is available.

6. A server-connected client device that provides recommendations, the client device comprising:

an input component that receives a user request;

a communication component that communicates the request to a server that:

stores an item attribute value representing the user's preference;

executes a recommendation engine to identify a recommendation based on the attribute value, wherein the recommendation engine determines the item from parsing a natural language expression of a person other than the user; and

communicates the recommendation to the client device; and

an output component that provides the recommendation to user.

7. The client device of claim 6 wherein the input component further receives information about a current environment of the client device, and wherein the communication component communicates the information about the current environment to the server, which stores a value of an environmental parameter in association with the item attribute value such that the recommendation is further based on the value of the environmental parameter and the information about the current environment.

8. The client device of claim 7 , wherein the environmental parameter is a type of activity.

9. The client device of claim 6 , wherein the client device syncs the user's preferences with the server when access is available.

10. A client-connected server for providing recommendations, the server comprising non-transitory computer readable medium for storing code, which when executed causes the server to:

receive, from a client, a request from a user;

identify, using a recommendation engine, a recommendation based on a stored item attribute value for an item representing a preference of the user, wherein the recommendation engine determines the item from parsing a natural language expression of a person other than the user; and

communicate the recommendation to the client.

11. The server of claim 10 further comprising software for receiving, from the client, information about the current environment of the client, wherein the server stores a value of an environmental parameter in association with the item attribute value, identifying the recommendation is further based on the value of the environmental parameter and the information about the current environment.

12. The server of claim 11 wherein the environmental parameter is a type of activity.

13. The server of claim 10 further comprising:

software for receiving, from the client, identification of a user; and

software to derive values of a user attribute vector by classifying the user based on attributes of other users in a user database,

wherein the recommendation is further based on values of the user attribute vector.

14. The server of claim 10 , wherein the client syncs the user's preferences with the server when access is available.

15. A server-connected client device that provides recommendations to a user and is in communication with a server that stores an item attribute value for an item representing the user's preference, executes a recommendation engine, which determines the item from parsing a natural language expression of a person other than the user, to identify a recommendation based on the attribute value, and communicates the recommendation to the client device, the client device comprising:

an input component that receives the user's request;

a communication component, in communication with the input component, that communicates the request to the server and receives the recommendation from the server; and

an output component, in communication with the communication component, that provides the recommendation to user.