US 8,620,917 B2Granted

Symantic framework for dynamically creating a program guide

AssigneeTelefonaktiebolaget L M Ericsson (publ)

InventorsSorin Marian Georgescu, Edoardo Gavita

Life of this patent

Estimating the term…

Loading patent…

Abstract

An application server includes a Semantic Analysis Core Service (SACS) function that communicates with a Semantic Analysis Client (SAC) in a Set Top Box (STB). The SACS groups programs available for rendering to a subscriber into program clusters. The SACS generates the program clusters based on a determined semantic similarity between the programs, and on parameters that indicate a subscriber's preference for certain program content. The program that are semantically similar to existing clusters within a predetermined viewing window are provided to the STB and output to the subscriber on a display as a program preference list or channel line-up. The STB also monitors the subscriber's interaction with the programs and calculates a preference score for each program indicating the subscriber's continuing, or waning, interest in a given program. The preference score is used to update the score of the program cluster to which the program belongs.

Claims29

Dependent claims

A method of clustering programs having similar content, the method comprising:

extracting terms from a description associated with a program that match attributes defined in a content ontology;
compiling a set of semantic features associated with the program based on the extracted terms;
for each of a plurality of program clusters, each cluster comprising a plurality of programs having similar content, calculating a semantic similarity value representing the semantic similarity between the set of semantic features associated with the program and a semantic feature set of the program cluster, the semantic similarity value being calculated as
sim
⁡
(
a
,
b
)
=
F
⁡
(
S
⁡
(
a
)
⁢
I
⁢
⁢
S
⁡
(
b
)
)
F
⁡
(
S
⁡
(
a
)
⁢
I
⁢
⁢
S
⁡
(
b
)
)
+
α
(
F
⁡
(
S
⁡
(
a
)
-
S
⁡
(
b
)
)
+
F
⁡
(
S
⁡
(
b
)
⁢
I
⁢
⁢
S
⁡
(
a
)
)
⁢
g
⁢
⁢
ⅇ
1
-
N

S
⁡
(
a
)
⁢
I
⁢
⁢
S
⁡
(
b
)

wherein:
sim(a,b) denotes the semantic similarity value;
F is a function that quantifies a semantic weight of the semantic features as specified by the content ontology;
S(a) denotes the semantic features associated with the program;
S(b) denotes the semantic features associated with the program cluster;
a is a constant;
e is Euler's constant;
and
N is a threshold indicating a number of relevant semantic features for the program cluster;
selecting the highest semantic similarity value resulting from a comparison of the program with all the program clusters in the plurality of program clusters;
and
if the highest semantic similarity value exceeds a predetermined threshold, adding the program to the associated cluster.

117 more elements

An application server configured to cluster programs having similar content, the server comprising:

memory configured to store a program description associated with a program;
a semantic analysis module communicatively connected to the memory and a content ontology database;
and
a programmable controller configured to:
extract terms from the program description that match attributes defined in the content ontology;
compile a set of semantic features associated with the program based on the extracted terms;
for each of a plurality of program clusters, each cluster comprising a plurality of programs having similar content, calculate a semantic similarity value representing the semantic similarity between the set of semantic features associated with the program and a semantic feature set of the program cluster, the semantic similarity value being calculated as
sim
⁡
(
a
,
b
)
=
F
⁡
(
S
⁡
(
a
)
⁢
I
⁢
⁢
S
⁡
(
b
)
)
F
⁡
(
S
⁡
(
a
)
⁢
I
⁢
⁢
S
⁡
(
b
)
)
+
α
(
F
⁡
(
S
⁡
(
a
)
-
S
⁡
(
b
)
)
+
F
⁡
(
S
⁡
(
b
)
⁢
I
⁢
⁢
S
⁡
(
a
)
)
⁢
g
⁢
⁢
ⅇ
1
-
N

S
⁡
(
a
)
⁢
I
⁢
⁢
S
⁡
(
b
)

wherein:
sim(a,b) denotes the semantic similarity value;
F is a function that quantifies a semantic weight of the semantic features as specified by the content ontology;
S(a) denotes the semantic features associated with the program;
S(b) denotes the semantic features associated with the program cluster;
a is a constant;
e is Euler's constant;
and
N is a threshold indicating a number of relevant semantic features for the program cluster;
select the highest semantic similarity value resulting from a comparison of the program with all the program clusters in the plurality of program clusters;
and
if the highest semantic similarity value exceeds a predetermined threshold, add the program to the associated cluster.

121 more elements

Dependencies are inferred from the claim text.