Method of providing selected content items to a user

1. A method of providing selected content items to a user, comprising the steps of:

A) identifying a first set of content items based on a request by the user, wherein first content metadata are pre-assigned to the content items of said first set;

B) automatically generating second content metadata for the content items of said first set on a basis of at least a first derivation rule, said derivation rule corresponding to an algorithm applied to at least the content items of said first set;

C) associating said second content metadata to the content items of said first set; and

D) providing a second set of selected content items from said first set on a basis of said first content metadata and said second content metadata,

wherein step D) is carried out based on a user profile of the user comprising a predictive model, and

wherein the method further comprises building the predictive model based on a feature vector and metadata, wherein the metadata comprises textual components and numerical components, and wherein the feature vector comprises metadata including at least one textual component from at least one of the first content metadata or the second content metadata,

wherein in step D) a ranking to each of the content items of said first set is provided on a basis of said second content metadata and of said predictive model so as to define said second set of content items based on said ranking.

2. The method according to claim 1 , wherein said algorithm is to be applied also to at least some of said first content metadata.

3. The method according to claim 1 , wherein step B) is carried out also on the basis of first context metadata relating to an interaction context of said selected content items in order to automatically generate second context metadata.

4. The method according to claim 1 , further comprising deriving said second content metadata from an algorithm applied to a plurality of content items.

5. The method according to claim 1 , wherein said ranking to each of the content items of said first set provided on the basis of said second content metadata and of said second context metadata and of said user profile, and to select said second set of content items based on said ranking.

6. The method according to claim 1 , wherein said second set of content items is provided as replies to corresponding requests by said user.

7. The method according to claim 1 , further comprising the step of associating a feedback from said user to at least a content item of said second set.

8. The method according to claim 7 , further comprising the step of recording at least a content item of said second set and a feedback from said user related to said at least a recorded content item.

9. The method according to claim 8 , comprising the step of recording at least part of said second content metadata used for selecting said at least a content item.

10. The method according to claims 8 , comprising the step of recording said user request.

11. The method according to claim 8 , comprising the step of building or updating the predictive model of said user based at least on said recorded content item and user feedback.

12. The method according to claims 11 , comprising the step of recording at least part of said second content metadata and the step of building or updating a predicative model of said user based at least on recorded second content metadata.

13. The method according to claim 12 , wherein said predictive model is built or updated through at least one machine learning algorithm applied to at least said second content metadata.

14. The method according to claim 13 , wherein said machine learning algorithm is applied also to at least some context metadata.

15. The method according to claim 1 , wherein the user profile comprises at least two predictive models.

16. The method according to claim 1 , wherein said second set of content items is provided by a service provider.

17. The method according to claim 1 , wherein said selected content items are provided through a telecommunication network.

18. A non-transitory computer readable medium encoded with a computer program product comprising software code portions, which when loaded into the memory of at least one computer and executed by the at least one computer, perform the method of claim 1 .

19. A method for providing a content-based service comprising:

collecting content requests from a user;

identifying a first set of content items based on said content requests by the user, wherein first content metadata are pre-assigned to the content items of said first set;

automatically generating second content metadata for the content items of said first set on a basis of at least a first derivation rule, said first derivation rule corresponding to an algorithm applied to at least the content items of said first set;

associating said second content metadata to the content items of said first set; and

providing a second set of selected content items from said first set on a basis of said first content metadata and said second content metadata,

wherein providing the second set is carried out based on a user profile of the user comprising a predictive model, and

wherein the method further comprises building the predictive model based on a feature vector and metadata, wherein the metadata comprises textual components and numerical components, and wherein the feature vector comprises metadata including at least one textual component from at least one of the first content metadata or the second content metadata,

wherein providing the second set comprises providing a ranking to each of the content items of said first set on a basis of said second content metadata and of said predictive model so as to define said second set of content items based on said ranking.