KR100832859B1 - Mobile web content service system and method - Google Patents

Abstract

The present invention relates to a mobile web content service system and a method thereof, wherein any web page requested from a mobile device is classified in block units by using a visual web page segmentation technique (VIPS), and a block of the classified web page is provided. Among them, filters other than the block containing the content of the webpage, remove the block, extract the title of the content from the block containing the content, and phrases having a frequency greater than or equal to a predetermined number of blocks containing the content as important phrases. Selects, summarizes the block containing the content using sentences containing the selected important phrase, compares the words included in the block containing the content with the words of interest of the prestored mobile device user, and the word of interest of the mobile device user. Contains content that contains words that are very similar to By allowing the title of the block to be displayed on the screen of the mobile device, the web content requested by the user of the mobile device is automatically converted to fit the mobile device, and the priority of the content list is provided to the individual users by learning using user interest information. It may be provided in a form modified to suit.

Mobile device, web content service

Description

Mobile web content service system and its method {MOBILE WEB CONTENTS SERVICE SYSTEM AND METHOD}

1 is a diagram showing a configuration of a mobile web content service system according to the present invention.

FIG. 2 is a diagram illustrating a flow of a method of classifying blocks of web content in the block classifier shown in FIG. 1.

3 is a diagram illustrating an example of title extraction of a title extraction unit shown in FIG. 1;

4 is a diagram illustrating an example of extracting important phrases from a content block of the summary unit illustrated in FIG. 1.

FIG. 5 is a diagram illustrating an example of a summary flow of important phrases of the summary unit illustrated in FIG. 1.

6 is a diagram illustrating an example of outputting a result of the filtering unit illustrated in FIG. 1 to a mobile device;

FIG. 7 is a diagram illustrating a mobile device performing results of the summary unit illustrated in FIG. 1;

8 is a diagram illustrating an example of a learning process of the learning unit illustrated in FIG. 1 on a mobile device screen;

The present invention relates to a mobile web content service system and a method thereof, and more particularly, automatically converts web content to a mobile device, and modifies the priority of a content list to a user by learning using user interest information. The present invention relates to a mobile web content service system and a method thereof.

Recently, in addition to the Internet, various mobile devices such as mobile phones and PDAs (Personal Digital Assistants) are widely used.

In addition, mobile device users want to see the rich variety of content available on their PCs on mobile devices. However, services are limited because of different performance, user preferences and network bandwidth for each mobile device.

Moreover, the mobile device has a limited screen size, and no module for providing various contents is provided. This eliminates this ancillary feature as mobile devices are designed to be lightweight. Therefore, a process called content adaptation is required for mobile users to receive web services widely used in general PCs without inconvenience.

Most content available on the web is not immediately available on mobile devices. This is because the web content has a large data size, or contains various types of content such as a format or sound file or multimedia information such as flash that cannot be represented on the mobile device.

Therefore, a method of converting various contents to be output on a mobile device has been issued.

However, when a user of a mobile device accesses a web server and selects web content, content of a mobile-only page created in the web server is provided to the user's mobile device.

This results in a problem that results in the opposite of what the user of the mobile device wants to preferentially provide information of interest in the list of mobile devices.

Therefore, the present invention was devised to solve the above problems, and automatically converts web content requested by the user of the mobile device to the mobile device, and prioritizes the content list through learning using user interest information. The purpose of the present invention is to provide a mobile web content service system and a method for providing a modified form to a user.

Mobile web content service system according to an aspect of the present invention for achieving the above object, block any web page requested from the mobile device using a visual web page segmentation technique (VIPS, Vision-based Page Segmentation Algorithm) A block classification unit classified into units; A filtering unit for filtering and removing blocks other than the block including the contents of the web page among the classified web page blocks; A title extraction unit which extracts a title of the content from the block including the content; A summary unit for selecting phrases having a frequency greater than or equal to a predetermined number in the block including the content as a critical phrase, and summarizing the block including the content using the sentences including the selected important phrase; And comparing the words included in the block including the content with the words of interest of the mobile device user previously stored, and the title of the block including the content including words having high similarity to the words of interest of the mobile device user is displayed on the screen of the mobile device. It includes a learning unit to be output.

The block classification unit outputs the web content divided into blocks on the screen of the mobile device.

The filtering unit classifies the classified blocks of the webpage into blocks other than blocks including the content of the webpage and blocks including the content of the webpage using a heuristic rule.

The summary unit selects, as an important phrase, a phrase having a frequency greater than or equal to a set number using a N-gram method in a block including the content.

The summary unit summarizes a block including content by removing sentences including a conjunction from among sentences containing the selected important phrase.

The learning unit may express words included in a block including the content and words of interest of a mobile device user previously stored on a vector using a vector model, and the words of interest of the user stored as the query of the vector model. The similarity with the block including the content is measured by cosine measurement, and the title of the block including the content containing the high similarity word is output on the top of the screen of the mobile device by the cosine measurement.

The mobile web content service method according to an aspect of the present invention for achieving the above object, block any web page requested from the mobile device using a visual web page segmentation technique (VIPS, Vision-based Page Segmentation Algorithm) Classifying the unit; Filtering and removing blocks other than the block including the content of the webpage from among the classified webpage blocks; Extracting a title of the content from the block including the content; Selecting phrases having a frequency greater than or equal to a predetermined number in the block including the contents as important phrases, and summarizing the blocks containing the content using sentences including the selected important phrases; And comparing the words included in the block including the content with the words of interest of the mobile device user previously stored, and the title of the block including the content including words having high similarity to the words of interest of the mobile device user is displayed on the screen of the mobile device. Causing the output to occur.

The classifying in block units may include outputting web content divided in block units on a screen of the mobile device.

The filtering and removing may include classifying the classified blocks of the webpage into blocks other than the blocks including the contents of the webpage and the blocks including the contents of the webpage using a heuristic rule.

The summarizing may include selecting, as an important phrase, a phrase having a frequency greater than or equal to a set number using an N-gram method in a block including the content.

The summarizing may include summarizing a block including content by removing sentences including conjunctions among sentences including the selected important phrase.

The step of outputting a title of a block including content including words with high similarity to words of interest of the mobile device user on the screen of the mobile device may include: words included in the block including the content; Expressing words of interest on a vector using a vector model, and measuring similarity with a block including the content by cosine measurement using the pre-stored user's word of interest as a query of the vector model; And outputting a title of a block including content having high similarity words to the screen of the mobile device through the cosine measurement.

Hereinafter, a mobile web content service system and a method thereof according to the present invention will be described in detail with reference to the accompanying drawings.

In addition, it should be noted that the same reference numerals are given to the same elements, although belonging to different drawings for convenience of understanding.

1 is a diagram showing the configuration of a mobile web content service system according to the present invention.

As shown in FIG. 1, the web content service system according to the present invention comprises a mobile device 10 and an adaptation server 20.

The adaptation server 20 includes a block classification unit 21, a filtering unit 22, a title extraction unit 23, a summary unit 24, a learning unit 25, and a database 26.

The block classification unit 21 classifies any web page requested by the user of the mobile device 10 in units of blocks by using a visual web page segmentation algorithm (VIPS).

The block classification unit 21 may output a web page classified in units of blocks on the screen of the mobile device 10.

The filtering unit 22 filters and removes blocks other than the blocks including the content of the web page using heuristic rules. In this case, blocks other than the block including the content of the web page may include a block including the content of the advertisement.

The title extractor 23 extracts a title of the content from the content block.

The summary unit 24 selects phrases having a frequency greater than or equal to a set number using a N-gram method in the content block as important phrases, and excludes sentences starting with conjunctions among sentences containing the selected important phrases. Summarize

The learner 25 compares the words included in the content block with the words of interest of the user of the mobile device 10 stored in the database 26, and compares the words of high interest with words of interest of the user of the mobile device 10. The title of the content block to be included is output at the top of the screen of the mobile device 10.

More specifically, the learning unit 25 expresses words included in a block including the content and words of interest of a mobile device user stored in the database 26 on a vector using a vector model, and the database 26 Using the user's words stored in the query as the query of the vector model, the similarity with the block including the content is measured by the cosine value, and the title of the block including the content including the high similarity word is determined by the cosine measurement. To be output to

FIG. 2 is a diagram illustrating a flow of a method of classifying blocks of web content in the block classification unit illustrated in FIG. 1.

As illustrated in FIG. 2, the block classifier determines that any web page requested by the user is a block including content if the content of the classified block through the web page segmentation technique (VIPS) is equal to or larger than the set size of the web page size (see FIG. 2). S20).

The block classification unit checks whether a link is included in the block including the content (S21).

If the block classification unit includes a link in the block including the content, the block classification unit checks whether there is another domain name in the included link (S22).

If there is another domain name in the included link, the block classification unit deletes the block (S23).

If there is no other domain name in the link included in step S22, the block classification unit checks whether the included link is at least half of the block including the content (S24).

If the text in the included link occupies more than half of the block including the content, the block classification unit checks whether the average length of the text in the link is greater than or equal to a set length (eg, "10") (S25).

If the included link is at least the length of the average link length is set, the block classification unit determines that the block as a navigation list block (S26).

In this case, the navigation list block represents a block in which links including content exist.

On the other hand, if the link included in step S25 is less than half of the block including the content, the block classification unit determines that the block as a navigation bar block (S27).

At this time, the navigation bar block represents a block having a menu.

On the other hand, if the link is not included in the block including the content in step S21, the block classification unit checks whether the number of words included in the block including the content is equal to or greater than the set number of words 100 (S29).

If the number of words included in the block including the content is greater than or equal to the set word number, the block classification unit determines the block including the content as the content block (S30).

In this case, the content block indicates that the block including the content is a block including the content.

On the other hand, if the number of words included in the block including the content is less than the set number of words, the block classification unit deletes the block (S31).

FIG. 3 is a diagram illustrating an example of title extraction of the title extraction unit shown in FIG. 1.

As shown in FIG. 3, the title extracting unit includes <title> tags 35 and <of tags constituting a web page to extract titles from the content block filtered by the filtering unit 24 shown in FIG. 1. meta name = "title"> tag 36 and <HI> tag 37 select duplicate contents as a title.

4 is a diagram illustrating an example of extracting an important phrase from a content block of the summary unit illustrated in FIG. 1.

As shown in FIG. 4, the summary unit summarizes the content of the content block output from the filtering unit 22 shown in FIG. 1 through the N-gram method.

In other words, the summary unit generates a phrase by combining tokens (eg, words) included in the content block, and generates a new phrase by combining the generated phrase with another token when the generated phrase exists over a set number of times.

This will be described in detail.

The summary unit combines the word 41 and the word 42 with respect to the words 41, 42, and 43 included in the web page 40 to generate a phrase, and is generated by the word 41 and the word 42. If there are more than the number of phrases set in the web page 40, the phrase is combined with the word 43 to create a new phrase. If the phrase generated by the word 41 and the word 42 is less than or equal to the number set in the web page 40, the word 42 and the word 43 are combined to generate a phrase, and the word 42 ) And the phrase generated by the word 43 exist over the set number of times.

Through this process, the summary unit extracts important phrases from the web page 40.

FIG. 5 is a diagram illustrating an example of a summary flow of important phrases of the summary unit illustrated in FIG. 1.

As shown in FIG. 5, the summary unit extracts a paragraph in which the important phrases 50 obtained by extracting the important phrases from the content block 51 exist as a summary candidate. 52 is obtained.

The summary unit obtains the summary 54 of the content block through an exception process 53 excluding the section beginning with the conjunction of the obtained summary candidates 52.

FIG. 6 is a diagram illustrating an example of outputting a result of the filtering unit illustrated in FIG. 1 to a mobile device.

In the mobile device, the navigation bar block 60 is shown as MENU, and the navigation list block 62 is not collecting the word of interest information of the mobile user when initially outputted to the mobile device. Print the title of the link that appears at the top. If there is a word of interest information of the mobile user, a link title of content of interest to the mobile user among the links is output to the mobile device 10. Meanwhile, parentheses of the navigation list block output to the mobile device 10 indicate the number of links in the navigation list block.

Finally, it can be seen that the content block 61 is displayed as “2.NASA Glenn” on the screen of the mobile device 10.

FIG. 7 is a diagram illustrating an example of an execution result of the summary unit illustrated in FIG. 1.

As shown in FIG. 7, a title 71 for content block 70 with content is displayed on the mobile device screen.

When the title 71 for the content block 70 displayed on the mobile device screen is selected in this way, the summary 72 for the title 71 is displayed on the mobile device screen.

If there is a selection of Full Story in the mobile device screen on which the summary 72 for the title 71 is displayed, the original text 73 for the title 71 is displayed on the mobile device screen.

By doing so, the user can see the text 71 divided according to the mobile device screen size.

FIG. 8 is a diagram illustrating an example of a learning process of the learning unit illustrated in FIG. 1 on a screen of a mobile device.

After the selection of the third navigation list block displayed on the screen 80 of the mobile device, if the user selects the contents related to "Afghan" among the links and views the original text, the word "Afghan" is stored in the user's word of interest. . The learning unit recognizes that the user is interested in "Afghan" and places the content at the top of the list.

In the third mobile device, it can be seen that the order of the second and third has been changed on the screen 82 of the first mobile device.

As described above, according to the mobile web content web service system and method thereof according to the present invention, the web content requested by the user of the mobile device is automatically converted to fit the mobile device, the content by learning using the user interest information The priority of the list can be provided in a form customized for each user.

Claims (14)

In the mobile web content service system, A block classification unit classifying any web page requested from the mobile device into blocks by using a visual web page segmentation technique (VIPS); A filtering unit for filtering and removing blocks other than the block including the contents of the web page among the classified web page blocks; A title extraction unit which extracts a title of the content from the block including the content; A summary unit for selecting phrases having a frequency greater than or equal to a predetermined number in the block including the content as a critical phrase, and summarizing the block including the content using the sentences including the selected important phrase; And The words included in the block including the content are compared with the words of interest of the mobile device user, and the title of the block including the content including words with high similarity to the words of interest of the mobile device user is displayed on the screen of the mobile device. Mobile web content service system comprising a learning unit to be. The method of claim 1, The block classification unit, Mobile web content service system for outputting the web content divided into blocks on the screen of the mobile device. The method of claim 1, The filtering unit, And classifying the blocks of the classified webpage into blocks including content of the webpage and blocks other than blocks containing the content of the webpage using heuristic rules. The method of claim 1, The summary section, A mobile web content service system for selecting phrases having a frequency greater than or equal to a predetermined number as an important phrase in a block including the content using an N-gram method. The method of claim 1, The summary section, The mobile web content service system of summarizing the block containing the content by removing the sentences including the conjunctions among the sentences containing the selected important phrases. The method of claim 1, The learning unit, Words included in the block including the content and words of interest of a mobile device user, which are previously stored, are expressed on a vector using a vector model, and the content of the word of interest is stored as a query of the vector model. And measure the similarity with the block by cosine measurement, and output the title of the block including the content containing the high similarity word on the screen of the mobile device by cosine measurement. In the mobile web content service system, A block classifier which classifies any web page requested from the mobile device into block units using a visual web page segmentation algorithm (VIPS), and a web page among the blocks of the classified web pages. A filtering unit for filtering and removing blocks other than the block containing the content of the content, a title extracting unit extracting a title of the content from the block containing the content, and a phrase having a frequency equal to or greater than a set number of frequencies in the block containing the content. Selects an important phrase, compares a summary section summarizing a block including content using sentences containing the selected important phrase, compares words included in the block containing the content with words of interest of a previously stored mobile device user, Cones that contain words that are most similar to words of interest to the mobile device user. Includes a learning unit for displaying the title of the block containing the content on the screen of the mobile device, The block classification unit outputs the web content divided into block units on the screen of the mobile device, and the filtering unit includes the blocks including the contents of the web page and the contents of the web page using heuristic rules. Are classified as blocks other than blocks, The summary unit selects a phrase having a frequency greater than or equal to a predetermined number using a N-gram method in a block including the content, and removes a sentence including a connection word from among sentences containing the selected important phrase and includes a content block. To summarize, The learning unit expresses the words included in the block including the content and the words of interest of the mobile device user previously stored on the vector using a vector model, and the contents of the content by using the prestored user's words of interest as a query of the vector model. The mobile web content service system for measuring the similarity with the block including a cosine measurement, and the title of the block including the content containing the high similarity words by the cosine measurement is displayed on the screen of the mobile device. In the mobile web content service method, Classifying any web page requested from the mobile device into units of blocks using a visual web page segmentation technique (VIPS); Filtering and removing blocks other than the block including the content of the webpage from among the classified webpage blocks; Extracting a title of the content from the block including the content; Selecting phrases having a frequency greater than or equal to a predetermined number in the block including the contents as important phrases, and summarizing the blocks containing the content using sentences including the selected important phrases; And The words included in the block including the content are compared with the words of interest of the mobile device user, and the title of the block including the content including words with high similarity to the words of interest of the mobile device user is displayed on the screen of the mobile device. Mobile Web content service method comprising the step of. The method of claim 8, The step of classifying the block unit, The mobile web content service method comprising the step of outputting the web content divided into blocks on the screen of the mobile device. The method of claim 8, Removing by filtering, And classifying the classified blocks of the webpage into blocks including content of the webpage and blocks other than blocks containing the content of the webpage using heuristic rules. The method of claim 8, Summarizing the step, The mobile web content service method comprising the step of selecting a phrase having an important frequency or more in the block containing the content by using an N-gram method as an important phrase. The method of claim 8, Summarizing the step, The mobile web content service method comprising the step of summarizing the block containing the content by removing the sentences including the connection among the sentences containing the selected important phrases. The method of claim 8, The step of outputting the title of the block including the content containing the words of high similarity to the words of interest of the mobile device user on the screen of the mobile device, Words included in the block including the content and words of interest of a mobile device user, which are previously stored, are expressed on a vector using a vector model, and the content of the word of interest is stored as a query of the vector model. Measuring similarity with the block by cosine measurement; Wow And outputting a title of a block including content having high similarity words to the screen of the mobile device by the cosine measurement. In the mobile web content service method, Classifying any web page requested from the mobile device into block units by using a visual web-based segmentation algorithm (VIPS), and the content of the web page among the classified web page blocks. Filtering and removing blocks other than the block including; and extracting a title of the content from the block including the content; selecting phrases having a frequency greater than or equal to a predetermined number of blocks including the content; Summarizing the block containing the content using sentences including the selected important phrase, comparing words included in the block including the content with interest words of the prestored mobile device user, and interest word of the mobile device user. Block containing content that contains words that are very similar to Sore throat comprising the step of: so that the output of the mobile device screen, The classifying of the block unit may include outputting the web content divided into block units on the screen of the mobile device, and the filtering and removing may include removing the blocks of the classified web page using a heuristic rule. Classifying the block into blocks other than the block containing the content and the block including the content of the web page, and the summarizing step includes a phrase having a frequency equal to or greater than a predetermined number using a N-gram method. Including the step of selecting, and the step of summarizing includes the step of summarizing the block containing the content by removing the sentences including the conjunctions among the sentences containing the selected important phrases, The step of outputting the title of the block including the content containing the words with high similarity to the words of interest of the mobile device user on the screen of the mobile device may include the words included in the block including the content and the words of interest of the mobile device user. Expressing them on a vector using a vector model, measuring the similarity with the block including the content by cosine measurement using the pre-stored user's word of interest as the query of the vector model, and the similarity by the cosine measurement. Causing the title of the block containing the content containing the high words to be output on the screen of the mobile device.

Images