System and method for adaptive text recommendation
View Patent ↗Network system provides a real-time adaptive recommendation set of documents with a high statistical measure of relevancy to the requestor device. The recommendation set is optimized based on analyzing text of documents of the interest set, categorizing these documents into clusters, extracting keywords representing the themes or concepts of documents in the clusters, and filtering a population of eligible documents accessible to the system utilizing site and or Internet-wide search engines. The system is either automatically or manually invoked and it develops and presents the recommendation set in real-time. The recommendation set may be presented as a greeting, notification, alert, HTML fragment, fax, voicemail, or automatic classification or routing of customer e-mail, personal e-mail, job postings, and offers for sale or exchange.
1. A method for adaptive text recommendation, the method comprising:
receiving a query submitted by a client;
computing a plurality of similarity scores between a plurality of documents that are of interest to the client, each of the plurality of similarity scores indicating similarity of words in a first document D 1 and words in a second document D 2 , and each similarity score being computed according to:
similarity
(
D
1
,
D
2
)
=
∑
w
∈
D
1
⋂
D
2
count
(
w
,
D
1
)
count
(
w
,
D
2
)
[
∑
w
∈
D
1
⋂
D
2
count
(
w
,
D
1
)
2
]
1
/
2
[
∑
w
∈
D
1
⋂
D
2
count
(
w
,
D
2
)
2
]
1
/
2
;
grouping the plurality of documents into a plurality of clusters based on the similarity scores;
constructing a recommended set by selecting one or more documents from the plurality of clusters, wherein the construction of the recommended set of documents further comprises calculating a relevance score of each document in the eligible set of documents;
selecting documents of the eligible set of documents with high relevance scores;
applying other selection criteria comprising popularity of the document in the eligible set of documents and client preference for the document in the eligible set of documents; and
presenting the recommended set to the client.
2. The adaptive text recommendation method of claim 1 wherein selecting one or more documents from the plurality of clusters includes:
extracting one or more keywords for each of the plurality of clusters;
computing a keyword score for each of the plurality of clusters;
selecting a plurality of eligible clusters based on the keyword score;
computing a relevance score for each document in the plurality of eligible clusters; and
selecting one or more documents based on the relevance score or both.
3. The adaptive text recommendation method of claim 2 wherein the keyword score is computed according to:
Keyword score( w,C )=log Frequency( w,C )−log Frequency( w ).
4. The adaptive text recommendation method of claim 2 , wherein the relevance score of a document D with respect to cluster C is computed according to:
relevance
(
D
,
C
)
=
∑
w
∈
keywords
(
C
)
count
(
w
,
D
)
[
∑
w
∈
keywords
(
C
)
count
(
w
,
D
)
2
]
1
/
2
.
5. The adaptive text recommendation method of claim 2 wherein the extracting keywords step utilizes a process that calculates the keyword score of the cluster and select keywords that maximizes the keyword score of the cluster.
6. The adaptive text recommendation method of claim 2 further comprising presenting the recommended set of documents suing a presentation technique that comprises sending an e-mail, displaying a greeting, displaying an HTML, fragment, sending a fax, sending a voicemail, sending a video alert, sending an audio alert, and transmitting a file representing the recommended set of documents.
7. The adaptive text recommendation method of claim 1 wherein the received query comprises a request from a requestor device enabled by an action of the client and a software request.
8. An adaptive text recommendation system comprising:
a processor configured to:
receiving a query submitted by a client;
computing a plurality of similarity scores between a plurality of documents that are of interest to the client, each of the plurality of similarity scores indicating similarity of words in a first document D 1 and words in a second document D 2 , and each similarity score being computed according to:
similarity
(
D
1
,
D
2
)
=
∑
w
∈
D
1
⋂
D
2
count
(
w
,
D
1
)
count
(
w
,
D
2
)
[
∑
w
∈
D
1
⋂
D
2
count
(
w
,
D
1
)
2
]
1
/
2
[
∑
w
∈
D
1
⋂
D
2
count
(
w
,
D
2
)
2
]
1
/
2
;
grouping the plurality of documents into a plurality of clusters based on the similarity scores;
constructing a recommended set by selecting one or more documents from the plurality of clusters, wherein the construction of the recommended set of documents further comprises calculating a relevance score of each document in the eligible set of documents;
selecting documents of the eligible set of documents with high relevance scores;
applying other selection criteria comprising popularity of the document in the eligible set of documents and client preference for the document in the eligible set of documents; and
presenting the recommended set to the client; and
a memory coupled to the processor, configure to provide the processor with instructions.
9. A computer storage medium storing the computer readable code for causing a computer system to execute the steps of an adaptive text recommendation system, the steps comprising:
receiving a query submitted by a client;
computing a plurality of similarity scores between a plurality of documents that are of interest to the client, each of the plurality of similarity scores indicating similarity of words in a first document D 1 and words in a second document D 2 , and each similarity score being computed according to:
similarity
(
D
1
,
D
2
)
=
∑
w
∈
D
1
⋂
D
2
count
(
w
,
D
1
)
count
(
w
,
D
2
)
[
∑
w
∈
D
1
⋂
D
2
count
(
w
,
D
1
)
2
]
1
/
2
[
∑
w
∈
D
1
⋂
D
2
count
(
w
,
D
2
)
2
]
1
/
2
;
grouping the plurality of documents into a plurality of clusters based on the similarity scores;
constructing a recommended set by selecting one or more documents from the plurality of clusters, wherein the construction of the recommended set of documents further comprises calculating a relevance score of each document in the eligible set of documents;
selecting documents of the eligible set of documents with high relevance scores;
applying other selection criteria comprising popularity of the document in the eligible set of documents and client preference for the document in the eligible set of documents; and
presenting the recommended set to the client.