Portable communication terminal for extracting subjects of interest to the user, and a method therefor

Patent metadata

Field	Value
Publication number	US-9323845-B2
Application number	US-201113577149-A
Country	US
Kind code	B2
Filing date	Jan 31, 2011
Priority date	Feb 3, 2010
Publication date	Apr 26, 2016
Grant date	Apr 26, 2016

Abstract

A portable communication device for extracting a user interest comprises a term vector generation unit for generating, based on types of text data stored in the portable communication device, a term vector representing each text data, a subject classification tree storage unit for storing a subject classification tree, which is a tree structure in which multiple nodes, each including at least one training data and representing a subject, are connected to one another, and a similarity calculation unit for calculating a similarity between the term vector and the training data for each node in the subject classification tree. The similarity calculation unit extracts a node name representing the user interest from the subject classification tree based on the similarity.

First claim

What is claimed is:

1. A portable communication device for extracting a user interest, the device comprising:
a term vector generation unit for generating, based on types of text data stored in the portable communication device, a term vector representing each text data;
a subject classification tree storage unit for storing a subject classification tree, which is a tree structure in which multiple nodes, each including at least one training data and representing a subject, are connected to one another;
a subject classification tree generation unit for generating the subject classification tree by processing open directory data;
a training data generation unit for generating the training data representing each directory based on text data information of a set of web sites included in the each directory of the open directory data;
a classification unit for mapping the training data to a directory included in the subject classification tree; and
a similarity calculation unit for calculating a similarity between the term vector and the training data for each node in the subject classification tree,
wherein the similarity calculation unit extracts a node name representing the user interest from the subject classification tree based on the similarity.

2. The portable communication device for extracting a user interest of claim 1 ,
wherein the term vector generation unit comprises:
a term extraction unit for extracting terms from the text data; and
a term weight calculation unit for calculating a term weight based on usage frequency of each of the terms used in the text data, and generation time of the text data containing the terms.

3. The portable communication device for extracting a user interest of claim 1 ,
wherein the similarity calculation unit calculates the similarity between the term vector and the training data included in each node of the subject classification tree, and
for each node of the subject classification tree, a parent node's similarity is calculated by summing up all its children's similarities.

4. The portable communication device for extracting a user interest of claim 1 ,
wherein the name of the node having the highest similarity in the subject classification tree is extracted as the user interest.

5. The portable communication device for extracting a user interest of claim 1 ,
wherein the text data are extracted from at least one of a text message, a file name, an e-mail, and a mobile web usage history generated in the portable communication device.

6. A method for extracting a user interest, the method comprising:
a step in which a term extraction unit extracts terms from text data stored in a portable communication device;
a step in which a term weight calculation unit calculates a term weight based on usage frequency of each of the terms used in the text data, and generation time of the text data containing the terms;
a step in which a term vector generation unit, based on types of text data stored in the portable communication device, generates a term vector representing each text data; and
a step in which an open directory data collection unit collects various open directories and information about web pages included in each of the directories;
a step in which a subject classification tree generation unit generates a subject classification tree by processing the collected directory data;
a step in which a training data generation unit generates a training data representing each directory based on text data information of a set of web sites included in the each directory of the collected directory data;
a stage in which a classification unit maps the training data to a directory included in the subject classification tree; and
a step in which a similarity calculation unit calculates a similarity between the term vector and the training data for each node included in a subject classification tree, which is a tree structure in which multiple nodes, each including at least one training data and representing a subject, are connected to one another;
wherein the similarity calculation unit extracts a node name representing the user interest from the subject classification tree based on the calculated similarity.

7. The method for extracting a user interest of claim 6 ,
wherein the similarity calculation unit, for each node of the subject classification tree, calculates a parent node's similarity by summing up all its children's similarities.

8. The method for extracting a user interest of claim 6 ,
wherein the similarity calculation unit extracts the name of the node having the highest similarity in the subject classification tree as the user interest.

Assignees

• Lee Sang Keun
• Ha Jong Woo
• Lee Jung Hyun
• Univ Korea Res & Bus Found

Inventors

• Lee Sang Keun
• Ha Jong Woo
• Lee Jung Hyun

Classifications

• G06F18/24137

  Distances to cluster centroïds · CPC title

• G06F18/24323

  Tree-organised classifiers · CPC title

• G06F16/284

  Relational databases · CPC title

• G06F16/353

  into predefined classes · CPC title

• G06Q30/02

  Marketing; Price estimation or determination; Fundraising · CPC title