KR20210023636A - Method and system for long-term relevant document clustering - Google Patents

Abstract

Disclosed are a method and a system for long-term highly relevant document clustering. The document clustering method comprises the steps of: collecting clusters clustered into similar documents on a time basis; and generating an issue cluster as a cluster group by merging the clusters based on the similarity between the clusters.

Description

Method and system for clustering documents with high long-term relevance {METHOD AND SYSTEM FOR LONG-TERM RELEVANT DOCUMENT CLUSTERING}

The following description relates to a technique for document clustering.

Recently, various studies have been conducted to systematically classify and cluster documents for intelligent information services such as information search and recommendation.

Clustering, that is, clustering, is a technique of data mining, which computes the similarity based on the attributes of multiple items and then clusters them.

Clustering is a process of dividing a given data set into several clusters that have similarities to each other, and data belonging to one cluster have similarities that are distinguished from data within different clusters.

In the field of information retrieval as a clustering method of such documents, various methods including a K-means clustering method and a clustering method using an ontology are being studied.

For example, Korean Patent Registration No. 10-1067819 (announced on September 27, 2011) discloses a technology for clustering documents using an ontology including the subject words of the document.

We provide a method and system for creating an issue cluster as a cluster group of related issue units in the mid- to long-term by merging short-term clusters with high correlation among short-term clusters created for each time period.

A document clustering method executed on a computer system, wherein the computer system comprises at least one processor configured to execute computer readable instructions contained in a memory, and the document clustering method comprises, by the at least one processor, time Collecting clusters clustered with similar documents based on the reference; And generating an issue cluster as a cluster group by merging the clusters based on the similarity between the clusters by the at least one processor.

According to an aspect, the generating may include merging the issue cluster and the cluster based on a similarity between the issue cluster and the cluster.

According to another aspect, the generating may include merging the issue cluster and the other issue cluster based on a similarity between the issue cluster and another issue cluster.

According to another aspect, in the step of merging the issue cluster and the other issue cluster, documents of another cluster may be merged into a cluster having a larger number of documents among the issue cluster and the other issue cluster.

According to another aspect, in the generating step, the issue cluster may be generated by comparing the similarity between vectors of the clusters and merging clusters having a vector similarity meeting a predetermined cluster merging criterion into one cluster. .

According to another aspect, when a document is added to the cluster, the generating may include calculating a vector of a cluster to which a document is added using a vector previously calculated for the cluster.

According to another aspect, the generating step includes calculating a vector of an issue cluster in which other clusters are merged using a vector previously calculated for the issue cluster when another cluster is merged with the issue cluster. can do.

According to another aspect, the generating may include: selecting the most similar cluster as a merge target by comparing each cluster with a cluster created after the corresponding cluster in a parallel processing method; And sequentially merging the clusters selected as the merging target in a serial processing method.

According to another aspect, the generating may further include excluding the merged cluster so that it is not selected again as the merge target.

According to another aspect, in the collecting step, a cluster clustered with similar documents may be collected based on a degree of similarity between documents.

There is provided a computer program stored in a non-transitory computer-readable recording medium for executing the document clustering method on the computer system.

A non-transitory computer-readable recording medium in which a program for executing the document clustering method is recorded on a computer is provided.

A computer system, comprising: at least one processor configured to execute computer-readable instructions contained in a memory, the at least one processor comprising: a cluster collection unit configured to collect clusters clustered with similar documents on a time basis; And a cluster generation unit generating an issue cluster as a cluster group by merging the clusters based on the similarity between the clusters.

According to an embodiment of the present invention, by merging short-term clusters generated for each time period based on similarity between clusters, an issue cluster may be generated as a cluster group of related issue units in the mid- to long-term.

According to an embodiment of the present invention, an issue cluster is created by merging clusters based on the similarity between clusters, so that documents of related issues can be effectively clustered even if the similarity between documents changes due to a change in subject or content over time.

According to an embodiment of the present invention, cluster merging is performed by calculating a vector value of a cluster to which a document has been added or a cluster merged with another cluster using the vector value of the previously calculated cluster, thereby dramatically reducing the amount of computation and It can provide stable clustering performance.

1 is a diagram showing an example of a network environment according to an embodiment of the present invention.
2 is a block diagram illustrating an internal configuration of an electronic device and a server according to an embodiment of the present invention.
3 is a diagram illustrating an example of components that may be included in a processor of a server according to an embodiment of the present invention.
4 is a flowchart illustrating an example of a method that a server can perform according to an embodiment of the present invention.
5 to 8 are exemplary diagrams showing an example of a process of merging clusters based on similarity between clusters in an embodiment of the present invention.
9 is a flowchart illustrating an example of a cluster merging process according to an embodiment of the present invention.
10 to 12 illustrate examples of an issue timeline interface screen using an issue cluster according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Embodiments of the present invention relate to a document clustering technique.

In the embodiments including those specifically disclosed in this specification, by merging short-term clusters generated for each time period based on the similarity between clusters, an issue cluster can be created as a cluster group of related issue units in the mid- to long-term, through which clustering quality And improve clustering performance.

1 is a diagram illustrating an example of a network environment according to an embodiment of the present invention. The network environment of FIG. 1 shows an example including a plurality of electronic devices 110, 120, 130, and 140, a plurality of servers 150 and 160, and a network 170. 1 is an example for explaining the invention, and the number of electronic devices or servers is not limited as in FIG. 1.

The plurality of electronic devices 110, 120, 130, and 140 may be a fixed terminal implemented as a computer system or a mobile terminal. Examples of the plurality of electronic devices 110, 120, 130 and 140 include AI speakers, smart phones, mobile phones, navigation systems, computers, notebook computers, digital broadcasting terminals, personal digital assistants (PDAs), and PMPs. Portable Multimedia Player), tablet PCs, game consoles, wearable devices, internet of things (IoT) devices, virtual reality (VR) devices, and augmented reality (AR) devices. As an example, in FIG. 1, the shape of a smartphone is shown as an example of the electronic device 110, but in the embodiments of the present invention, the electronic device 110 is substantially different through the network 170 using a wireless or wired communication method. It may mean one of various physical computer systems capable of communicating with the electronic devices 120, 130, 140 and/or the servers 150 and 160.

The communication method is not limited, and not only a communication method using a communication network (for example, a mobile communication network, wired Internet, wireless Internet, broadcasting network, satellite network, etc.) that the network 170 may include, but also short-range wireless communication between devices is included. I can. For example, the network 170 includes a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), and a broadband network (BBN). And any one or more of networks such as the Internet. In addition, the network 170 may include any one or more of a network topology including a bus network, a star network, a ring network, a mesh network, a star-bus network, a tree or a hierarchical network, etc. Not limited.

Each of the servers 150 and 160 is a computer device or a plurality of computers that communicates with a plurality of electronic devices 110, 120, 130, and 140 through a network 170 to provide commands, codes, files, contents, services, etc. It can be implemented with devices. For example, the server 150 may be a system that provides a first service to a plurality of electronic devices 110, 120, 130, and 140 accessed through the network 170, and the server 160 is also a network ( It may be a system that provides a second service to a plurality of electronic devices 110, 120, 130, and 140 accessed through 170). As a more specific example, the server 150 provides a service (for example, a news service, etc.) targeted by the application through an application as a computer program installed and driven on a plurality of electronic devices 110, 120, 130, and 140. As a first service, it may be provided to a plurality of electronic devices 110, 120, 130, and 140. As another example, the server 160 may provide a service for distributing the above-described application installation and running file to the plurality of electronic devices 110, 120, 130, and 140 as the second service.

2 is a block diagram illustrating an internal configuration of an electronic device and a server according to an embodiment of the present invention. In FIG. 2, internal configurations of the electronic device 110 and the server 150 are described as examples of the electronic device. In addition, other electronic devices 120, 130, 140 or server 160 may also have the same or similar internal configuration as the electronic device 110 or server 150 described above.

The electronic device 110 and the server 150 may include memories 211 and 221, processors 212 and 222, communication modules 213 and 223, and input/output interfaces 214 and 224. The memories 211 and 221 are non-transitory computer-readable recording media, such as random access memory (RAM), read only memory (ROM), disk drive, solid state drive (SSD), flash memory, and the like. It may include a permanent mass storage device. Here, a non-destructive mass storage device such as a ROM, SSD, flash memory, disk drive, etc. may be included in the electronic device 110 or the server 150 as a separate permanent storage device separated from the memories 211 and 221. In addition, the memories 211 and 221 include an operating system and at least one program code (for example, a browser installed and driven in the electronic device 110 or a code for an application installed in the electronic device 110 to provide a specific service). Can be saved. These software components may be loaded from a computer-readable recording medium separate from the memories 211 and 221. Such a separate computer-readable recording medium may include a computer-readable recording medium such as a floppy drive, a disk, a tape, a DVD/CD-ROM drive, and a memory card. In another embodiment, software components may be loaded into the memories 211 and 221 through communication modules 213 and 223 other than a computer-readable recording medium. For example, at least one program is a computer program installed by files provided by a file distribution system (for example, the server 160 described above) that distributes the installation files of developers or applications through the network 170 It may be loaded into the memories 211 and 221 based on (for example, the above-described application).

The processors 212 and 222 may be configured to process instructions of a computer program by performing basic arithmetic, logic, and input/output operations. Instructions may be provided to the processors 212 and 222 by the memories 211 and 221 or the communication modules 213 and 223. For example, the processors 212 and 222 may be configured to execute a command received according to a program code stored in a recording device such as the memories 211 and 221.

The communication modules 213 and 223 may provide a function for the electronic device 110 and the server 150 to communicate with each other through the network 170, and the electronic device 110 and/or the server 150 A function for communicating with an electronic device (for example, the electronic device 120) or another server (for example, the server 160) may be provided. As an example, a request generated by the processor 212 of the electronic device 110 according to a program code stored in a recording device such as a memory 211 is sent to the server 150 through the network 170 under the control of the communication module 213. ) Can be delivered. Conversely, control signals, commands, contents, files, etc. provided under the control of the processor 222 of the server 150 are transmitted through the communication module 223 and the network 170 to the communication module 213 of the electronic device 110. ) Through the electronic device 110. For example, a control signal, command, content, file, etc. of the server 150 received through the communication module 213 may be transmitted to the processor 212 or the memory 211, and the content or file may be transmitted to an electronic device ( 110) may be stored as a storage medium (permanent storage device described above) that may further include.

The input/output interface 214 may be a means for an interface with the input/output device 215. For example, the input device may include a device such as a keyboard, a mouse, a microphone, and a camera, and the output device may include a device such as a display, a speaker, and a haptic feedback device. As another example, the input/output interface 214 may be a means for interfacing with a device in which input and output functions are integrated into one, such as a touch screen. The input/output device 215 may be configured with the electronic device 110 and one device. In addition, the input/output interface 224 of the server 150 may be a means for interface with a device (not shown) for input or output that is connected to the server 150 or that the server 150 may include. As a more specific example, when the processor 212 of the electronic device 110 processes a command of a computer program loaded in the memory 211, a service configured by using data provided by the server 150 or the electronic device 120 A screen or content may be displayed on the display through the input/output interface 214.

In addition, in other embodiments, the electronic device 110 and the server 150 may include more components than those of FIG. 2. However, there is no need to clearly show most of the prior art components. For example, the electronic device 110 may be implemented to include at least some of the input/output devices 215 described above, or may further include other components such as a transceiver, a camera, various sensors, and a database. As a more specific example, when the electronic device 110 is an AI speaker, various sensors, camera modules, various physical buttons, buttons using a touch panel, input/output ports, and various vibrators for vibration are generally included in the AI speaker. Components may be implemented to be further included in the electronic device 110.

Hereinafter, a specific embodiment of a method and system for document clustering with high long-term correlation will be described.

In the present specification, a document may mean an information unit that is an object of a search or recommendation on the Internet.

In this embodiment, the article (news) provided by the news service will be described as a representative example of the document, but the object of clustering is not limited to the article, and can be extended to all types of documents provided as information units on the Internet.

Document clustering technology is used to effectively classify and deliver many articles by subject, and a method of exposing articles of a specific cluster to the upper part of the service screen through a predetermined algorithm depending on the size or importance of the clustered article cluster.

Such a clustering technique clusters articles with similar contents, and the quality is high when mainly used for the latest articles, but the clustering quality decreases as the time range increases. In addition, there is a limit to clustering of suggestive articles whose subject changes over time, and there is a problem that the processing time increases rapidly as the number of articles increases.

Embodiments of the present invention are intended to create an issue cluster as a cluster group of related issue units in the medium to long term by merging short-term clusters generated for each time period.

3 is a block diagram showing an example of components that may be included in a processor of a server according to an embodiment of the present invention, and FIG. 4 is an example of a method that can be performed by a server according to an embodiment of the present invention. Is a flow chart showing.

The server 150 according to the present embodiment may serve as a news service platform that provides articles. In particular, the server 150 may effectively cluster and provide articles corresponding to issues that have persisted in the mid to long term.

The processor 222 of the server 150 is a component for performing the document clustering method according to FIG. 4, and as shown in FIG. 3, the cluster collection unit 310, the cluster generation unit 320, and the cluster exposure unit It may include (330). Depending on the embodiment, components of the processor 222 may be selectively included or excluded from the processor 222. In addition, according to an embodiment, components of the processor 222 may be separated or merged to express the function of the processor 222.

Components of the processor 222 and the processor 222 may control the server 150 to perform steps S410 to S450 included in the document clustering method of FIG. 4. For example, the processor 222 and the components of the processor 222 may be implemented to execute an instruction according to the code of the operating system included in the memory 221 and the code of at least one program.

Here, the components of the processor 222 may be expressions of different functions performed by the processor 222 according to an instruction provided by the program code stored in the server 150. For example, the cluster collection unit 310 may be used as a functional representation of the processor 222 that controls the server 150 according to the above-described command so that the server 150 collects a short-term cluster.

The processor 222 may read necessary commands from the memory 221 loaded with commands related to control of the server 150. In this case, the read command may include a command for controlling the processor 222 to execute steps S410 to S450 to be described later. The steps S410 to S450 to be described later may be performed in an order different from the order shown in FIG. 4, and some of the steps S410 to S450 may be omitted or an additional process may be further included.

Hereinafter, a cluster that is grouped by articles with high relevance on a time basis is called a'short-term cluster', and when a single issue unit is created by grouping into clusters with high relevance, the cluster group corresponding to the issue unit is called'issue cluster'. I will name it.

Referring to FIG. 4, in step S410, the cluster collection unit 310 may collect a short-term cluster clustered with similar articles based on the similarity between articles. The cluster collection unit 310 may collect a short-term cluster generated by clustering articles for a recent predetermined time. An article clustering method for generating a short-term cluster may use at least one of widely used clustering techniques. In this case, a unique identifier (ID) may be assigned to each short-term cluster during the cluster creation process. In the case of a cluster that was previously created and existed during the article clustering operation, the previously assigned identifier may be maintained.

For example, referring to FIG. 5, when the degree of similarity between vectors of different articles 50 is greater than or equal to a predetermined level, the cluster collection unit 310 may form a short-term cluster 501 by grouping them into one cluster. In this case, the vector of each article 50 may mean a vector average of words included in the article 50. The cluster collection unit 310 may group similar articles for the last 36 hours into a short-term cluster 501. Even if an identifier (ID) previously assigned to the short-term cluster 501 is maintained, articles that have passed 36 hours are deleted from the cluster, and only articles within the last 36 hours may be exposed as a service target. In the exemplary embodiment of the present invention, in collecting the short-term cluster 501, articles deleted over time in the short-term cluster 501 may also be accumulated and collected as an existing identifier (ID).

Referring to FIG. 4, in step S420, the cluster generation unit 320 may generate an issue cluster by merging short-term clusters based on the similarity between clusters. Some of the short-term clusters consist of articles on similar topics, but they cannot be grouped into one cluster or exist in different clusters due to the time difference between articles. To solve this problem, the cluster generation unit 320 may group short-term clusters into one issue cluster when short-term clusters meet the cluster merging criteria, that is, when the similarity between short-term clusters is greater than or equal to a certain level.

For example, referring to FIG. 6, the cluster generation unit 320 creates an issue cluster 602 by merging into one cluster when the similarity between vectors of the short-term cluster 501 among the short-term cluster 501 is greater than or equal to a certain level. I can. The vector of the short-term cluster 501 may mean a vector average of articles included in the corresponding cluster.

The degree of similarity between clusters is applied as a cluster merging criterion, but as an example, clusters having at least one of an Euclid value and a cosine value in a predetermined range may be grouped into one cluster by comparing cluster vectors. For example, when the cosine similarity between vectors of two clusters is 0.98 or more, and/or when the Euclidean similarity between vectors of two clusters is 0.71 or more, the two clusters may be grouped into one cluster. In this case, the degree of similarity between vectors may be defined as <1/(1+distance between two vectors)>.

To find the vector value of the cluster, when an article is added to the cluster, the method of obtaining the vectors of all articles and obtaining a new average is applied. In this case, since it involves a large number of API calls and a large number of calculations. As the number of articles increases, the amount of calculation also increases.

In this embodiment, when article n is added to the short-term cluster c, a new vector v'may be obtained at once using the vector v previously calculated for the short-term cluster c, which may be defined as Equation 1.

[Equation 1]

v'=(v _n +(v×m _c ))/(1+ m _c )

Here, v _n denotes the vector value of the added article n, and m _c denotes the number of existing articles included in the short-term cluster c.

Through this method, it is possible to quickly and easily obtain a new vector value of a short-term cluster to which an article has been added using the previously calculated vector value of the short-term cluster.

Referring to FIG. 4, in step S430, the cluster generation unit 320 may merge a newly created short-term cluster with an issue cluster created in step S420 based on the similarity between clusters.

7, the cluster generation unit 320 compares the vector of the issue cluster 602 generated in step S420 with the vector of the newly generated short-term cluster 501' to meet the cluster merging criteria. In the case of having a vector similarity, the short-term cluster 501 ′ may be included in the issue cluster 602. In this case, the vector of the issue cluster may mean a vector average of short-term clusters included in the corresponding cluster, and the merging criterion between the short-term cluster and the issue cluster is the same as the merging criterion between the short-term clusters described above.

Referring to FIG. 4, in step S440, the cluster generation unit 320 may merge previously generated issue clusters with the issue clusters generated in step S430 based on the similarity between clusters. The merging criterion between the issue cluster and the issue cluster is the same as the merging criterion between the short-term clusters described above.

As shown in FIG. 8, the cluster generation unit 320 compares the vector of the issue cluster generated through merging of short-term clusters with the vector of the previously generated issue cluster 602' to match the vector of the cluster merging criterion. If they have similarity, they may be merged into one issue cluster 602. In other words, the cluster generation unit 320 may compare vectors between issue clusters with each other and group them into one issue cluster 602 if it meets the cluster merging criteria. In this case, the cluster generation unit 320 may compare the number of articles of the issue cluster meeting the cluster merging criteria and merge the articles of the remaining issue clusters into an issue cluster having a large number of articles.

In this embodiment, when a new short-term cluster c is added to the issue cluster C, or a small issue cluster C'(with a smaller number of articles) is merged into a larger issue cluster C (with a larger number of articles), the issue cluster C is previously Using the calculated vector V, a new vector V'can be obtained at once, which can be defined as in Equation 2 or 3.

[Equation 2]

V'=((v _c ×m _c )+(V×m _C ))/(m _c +m _C )

Here, v _c is the vector value of the short-term cluster c, m _c is the number of articles included in the short-term cluster c, and m _C is the number of existing articles included in the issue cluster C.

[Equation 3]

V'=((v _C' ×m _C' )+(V×m _C ))/(m _C' +m _C )

Here, v _{C 'is} smaller artifact cluster C' vector value, m _{C 'is} smaller artifact cluster C' the number of articles contained in, _C m is the number of existing articles in issues cluster C.

Through this method, it is possible to quickly and easily obtain a new vector value of an issue cluster in which a short-term cluster or a small issue cluster is merged using the previously calculated vector value of the issue cluster.

Compared to obtaining the vector average of all articles each time for a cluster, a new vector of a cluster to which articles are added or merged with other clusters can be obtained at a rapid speed by using a vector value calculated in advance in the previous process.

The process of generating the issue cluster (S410 to S440) described above is exemplary and is not limited thereto.

And, assuming that there are short-term clusters or issue clusters of A, B, C, and D, when AB, AC, BC, and CD satisfy the cluster merging condition in the inter-cluster merging process (S420 to S440), merging is performed at once. When proceeding, A, B, C, and D are all merged into one cluster.

However, when clusters meeting the cluster merging conditions are merged at once, the following problems may occur.

First, as a result of the merger of clusters A and B, the vector distance between the vector of the merged cluster AB and the cluster D may become distant, so that the cluster merge condition may not be satisfied. As the vector distance becomes closer, the cluster merging condition may be met.

Second, when the cluster merging process is performed in parallel, clusters A and B may be merged into cluster B, and clusters B and C may be merged into cluster B.

An example of a cluster merging method for solving these two problems is shown in FIG. 9.

Referring to FIG. 9, the cluster generation unit 320 compares all clusters with clusters created after the corresponding cluster for all clusters, and merges only the cluster (short-term cluster or issue cluster) with the closest vector distance among them. It is selected as (S901). The above-described step (S901) proceeds in a parallel processing method.

The cluster generation unit 320 merges the clusters selected as a merge target in step S901 one by one through a serial processing method (S902). A cluster that has already been merged may be ignored even if it is selected as a merge target again in the next process, and may be excluded so that it is not selected as a merge target after the merge.

The cluster generation unit 320 determines whether the number of clusters to be merged is 0 as a result of performing step (S902) (S903), and if there is one or more clusters to be merged, the step S901 again targets all clusters including the merged clusters. Repeat from ).

The cluster generation unit 320 repeats the above processes (S901 to S902) until the number of clusters to be merged becomes 0, and ends cluster merging when the number of clusters to be merged becomes 0 (S904).

In the present embodiment, the cluster merging can be performed stably and quickly by performing parallel processing of a part of a large amount of computation without affecting the results before and after the cluster merging.

The cluster merging method can utilize various features representing similarity, such as the existence of the same article, vector Euclidean, and vector cosine, and at least one of the widely used merging methods can be used. It can be selectively applied according to requirements such as number, number of clusters, similarity or relevance of articles in the cluster.

In this embodiment, by merging volatile cluster data and fragmented cluster data in a service providing articles, an issue cluster can be created as a cluster group of related issue units in the medium to long term.

In the past, in the case of clustering articles collected for a long period of time, articles with similar texts are grouped around the text, resulting in similar document bundles, but in the present invention, clusters grouped by time are merged based on similarity between clusters. A cluster group can be created so that the content changes in the document, but in the context, a group of documents with common issues can be created. By creating an issue cluster by merging clusters based on similarity between clusters, articles that are difficult to merge with similarity between articles can be grouped into one cluster with common issues.

Referring to FIG. 4, in step S450, the cluster exposure unit 330 may expose articles of the issue cluster finally merged in step S440 through an article providing service. The cluster exposure unit 330 may provide a timeline for articles included in the issue cluster among articles on the Internet as an article timeline using an issue cluster. In other words, the cluster exposure unit 330 may expose articles included in the issue cluster in a time series by using the issue cluster.

The cluster exposure unit 330 may expose at least one issue cluster meeting a predetermined exposure condition among issue clusters to a service. As an example, the cluster exposure unit 330 targets an issue cluster in which an article specified by a content provider (eg, a media company) or an article specified by a user's selection or setting of a service is present. Can be determined by In other words, the cluster exposure unit 330 may expose an issue timeline on a service page for an issue cluster corresponding to a major issue handled by a media company or an issue of interest to an individual user. As another example, the cluster exposure unit 330 may determine an issue cluster in which at least one of the number of documents, the elapsed time after clustering, and the number of comments corresponds to a predetermined condition as a service exposure target. The exposure condition of the issue cluster is, for example, if the first clustered article among the articles in the cluster lasts more than 2 days, or if the total number of articles in the cluster is 200 or more, articles with 100 or more comments per day It may include three or more cases, and if the conditions are more than two days. It is also possible to determine the target of service exposure by combining media issues or personal issues with the exposure conditions of the issue cluster described above.

An example of an issue timeline interface screen using the above issue cluster is shown in FIGS. 10 to 12.

Referring to FIG. 10, the cluster exposure unit 330 may expose an issue cluster, that is, an issue timeline screen 1000 for displaying articles with high correlation in a mid- to long-term, through a news service. The issue timeline screen 1000 is a time area 1010 showing the number of articles per unit time (e.g., 1 day) included in the issue cluster in a graph form, and articles showing a list of articles included in the issue cluster divided by unit time The region 1020 may be included.

The issue timeline screen 1000 may identify articles with high correlation in the mid to long term at a glance, and may provide a navigation function for checking articles in a structure that is organically connected between the time domain 1010 and the article area 1020.

A partial area of the issue timeline screen 1000, for example, at the top, may include information about a period in which an issue cluster was created 1001, an issue title 1002 related to an issue cluster, and the like. The issue title 1002 is created by using the title or tag of at least one article included in the issue cluster, for example, the title of the article with the largest number of comments may be displayed as the issue title 1002, or An issue title 1002 may be generated and displayed as a combination of keywords that mainly appear in the titles or tags of articles.

The time domain 1010 contains time information of each unit time included in the issue cluster and the number of articles. For example, time information is displayed on one axis and the number of articles is displayed on the other axis in the form of a bar graph. Can be. In this case, display elements (eg, color, brightness, etc.) of the graph bar may be classified and displayed according to the number of articles. For example, as the number of articles increases, the color of the graph bar may be darker, and as the number of articles decreases, the color of the graph bar may be displayed lighter.

In the article area 1020, a list of articles by unit time merged into one issue cluster may be sorted according to a certain criterion, such as by date and number of articles. For example, the most recent cluster is exposed at the top or the most articles The cluster containing is can be exposed at the top level.

Referring to FIG. 11, in the issue timeline screen 1000, the time area 1010 can be scrolled 1101 in one direction (eg, left and right), and the article area 1020 is the same as the time area 1010. Alternatively, the interface may be configured to enable the scroll 1102 in a different direction (eg, vertical direction).

As the issue cluster is exposed to the service, the user can check the number of articles per unit time within the entire clustered period after the issue appeared through scrolling 1101 on the time domain 1010 of the issue cluster. , You can check the change of the issue over time, such as when it became the most issue, at a glance.

In addition, through the scroll 1102 on the article area 1020, a list of articles by unit time included in the issue cluster can be displayed in a predetermined sorting order, and a user can check the articles clustered by unit time. In the article area 1020, only some articles of a certain number (eg, 3) are displayed for each unit time, and when a separate request (eg, view more, view all, etc.) for the article list on a specific date is input by the user, the entire article is displayed. Can show.

When the article area 1020 is scrolled 1102, the time area 1010 may be automatically scrolled according to the screen exposure of the article list. For example, according to the scroll 1102 for the article area 1020, the unit time graph bar of the article list exposed in a predetermined reference range (eg, center line, etc.) of the screen in the time area 1010 is automatically scrolled to the center. It can be displayed to be distinguished from other unit time graph bars.

Referring to FIG. 12, in the process of checking an article list for each unit time through a scroll 1102 for the article area 1020, the time area 1010 may be simplified compared to the initial state (FIG. 10 ). For example, if the scroll 1102 for the article area 1020 lasts for a certain period of time (eg, 1 second) or more, the size of the time area 1010 is reduced or the graph bar of the time area 1010 is displayed in a reduced form. Can be. When the article list arranged at the top according to the scroll 1102 of the article area 1020 is exposed on the screen again, the time area 1010 may be restored to an initial state (FIG. 10).

In addition, whether the time region 1010 is exposed, that is, an exposed state or a hidden state may be selectively applied according to the direction of the scroll 1102 with respect to the article region 1020. For example, when the article area 1020 is scrolled downward, the time area 1010 is hidden, whereas when the article area 1020 is scrolled upward, the time area 1010 may be exposed again. have.

Likewise, when the time area 1010 is scrolled 1101, the article area 1020 may be automatically scrolled according to the screen exposure of the unit time graph bar. In other words, the time area 1010 and the article area 1020 are connected to each other, so that another area may be automatically scrolled according to the scrolling of one area.

As another example, when the article area 1020 is scrolled 1102, the time area 1010 is automatically scrolled, whereas when the time area 1010 is scrolled 1101, the article area 1020 is moved from an existing position. It is also possible to maintain a fixed state, of course, and vice versa.

When the interface is configured so that the article area 1020 is not automatically scrolled and remains fixed when scrolling 1101 for the time area 1010, a graph of a specific date in the time area 1010, for example, a graph of July 26 When the bar is selected, the article list of July 26 may be automatically scrolled and displayed in the article area 1020. In the time area 1010, a graph bar of a selected date may be displayed to be distinguished from another unit time graph bar. When a graph bar of a specific date is selected in the initial state of the time domain 1010 (FIG. 10 ), a list of articles of the selected date may be displayed in the article area 1020 and the time area 1010 may be simplified and displayed.

As described above, according to embodiments of the present invention, by merging short-term clusters generated for each time period based on the similarity between clusters, an issue cluster may be created as a cluster group of related issue units in the medium to long term. By creating an issue cluster by merging clusters based on the similarity between clusters, it is possible to effectively cluster articles of related issues even if the similarity between articles changes due to changes in the subject or content over time. By performing cluster merging by calculating the vector value of the cluster to which the article was added or the cluster merged with other clusters using the previously calculated vector value of the cluster, it is possible to provide fast and stable clustering performance by drastically reducing the amount of computation. .

In addition, according to an embodiment of the present invention, it is possible to provide an article timeline capable of showing articles with high correlation in a medium to long term in a time series using an issue cluster. By displaying the display area for unit time according to the issue cluster and articles clustered by each unit time, articles with high correlation for a long time can be identified at a glance, and a navigation function for checking articles in an organically connected structure between the time domain and the article area By providing a user convenience can be improved.

The apparatus described above may be implemented as a hardware component, a software component, and/or a combination of a hardware component and a software component. For example, the devices and components described in the embodiments include a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), and a programmable gate array (PLU). It may be implemented using one or more general purpose or special purpose computers, such as a logic unit), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications executed on the operating system. Further, the processing device may access, store, manipulate, process, and generate data in response to the execution of software. For the convenience of understanding, although it is sometimes described that one processing device is used, one of ordinary skill in the art, the processing device is a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that it may include. For example, the processing device may include a plurality of processors or one processor and one controller. In addition, other processing configurations are possible, such as a parallel processor.

The software may include a computer program, code, instructions, or a combination of one or more of these, configuring the processing unit to operate as desired or processed independently or collectively. You can command the device. Software and/or data may be embodyed in any type of machine, component, physical device, computer storage medium or device to be interpreted by the processing device or to provide instructions or data to the processing device. have. The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. In this case, the medium may be one that continuously stores a program executable by a computer, or temporarily stores a program for execution or download. Further, the medium may be a variety of recording means or storage means in a form in which a single or several pieces of hardware are combined, but is not limited to a medium directly connected to a computer system, but may be distributed on a network. Examples of media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, And ROM, RAM, flash memory, and the like may be configured to store program instructions. In addition, examples of other media include an app store that distributes applications, a site that supplies or distributes various software, and a recording medium or a storage medium managed by a server.

Although the embodiments have been described by the limited embodiments and drawings as described above, various modifications and variations can be made from the above description to those of ordinary skill in the art. For example, the described techniques are performed in a different order from the described method, and/or components such as systems, structures, devices, circuits, etc. described are combined or combined in a form different from the described method, or other components Alternatively, even if substituted or substituted by an equivalent, an appropriate result can be achieved.

Therefore, other implementations, other embodiments, and those equivalent to the claims also fall within the scope of the claims to be described later.

Claims (20)

In the document clustering method executed in a computer system,
The computer system includes at least one processor configured to execute computer readable instructions contained in a memory,
The document clustering method,
Collecting, by the at least one processor, a cluster clustered with similar documents on a time basis; And
Creating an issue cluster as a cluster group by merging the clusters based on the similarity between the clusters by the at least one processor
Document clustering method comprising a. The method of claim 1,
The generating step,
Merging the issue cluster and the cluster based on the similarity between the issue cluster and the cluster
Document clustering method comprising a. The method of claim 1,
The generating step,
Merging the issue cluster and the other issue cluster based on the similarity between the issue cluster and another issue cluster
Document clustering method comprising a. The method of claim 3,
The step of merging the issue cluster with the other issue cluster,
Merging documents of another cluster into a cluster having a larger number of documents among the issue cluster and the other issue cluster
Document clustering method, characterized in that. The method of claim 1,
The generating step,
Comparing the similarity between vectors of the clusters and merging clusters having a vector similarity meeting a predetermined cluster merging criterion into one cluster to create the issue cluster
Document clustering method, characterized in that. The method of claim 5,
The generating step,
When a document is added to the cluster, calculating a vector of the cluster to which the document is added using the vector previously calculated for the cluster
Document clustering method comprising a. The method of claim 5,
The generating step,
When another cluster is merged into the issue cluster, calculating a vector of the issue cluster in which the other clusters are merged using the vector previously calculated for the issue cluster
Document clustering method comprising a. The method of claim 1,
The generating step,
In a parallel processing method, comparing each cluster with a cluster created after the corresponding cluster and selecting the most similar cluster as a merge target; And
In a serial processing method, sequentially merging the clusters selected as the merging target
Document clustering method comprising a. The method of claim 8,
The generating step,
Excluding the merged cluster so that it is not selected again as the merge target
Document clustering method further comprising a. The method of claim 1,
The collecting step,
Collecting clusters clustered with similar documents based on the degree of similarity between documents
Document clustering method, characterized in that. A computer program stored in a non-transitory computer-readable recording medium for executing the document clustering method of claim 1 on the computer system. A non-transitory computer-readable recording medium in which a program for executing the document clustering method of any one of claims 1 to 10 on a computer is recorded. In a computer system,
At least one processor configured to execute computer readable instructions contained in memory
Including,
The at least one processor,
A cluster collection unit for collecting clusters clustered with similar documents based on time; And
A cluster generation unit that creates an issue cluster as a cluster group by merging the clusters based on the similarity between the clusters
Computer system comprising a. The method of claim 13,
The cluster generation unit,
Merging the issue cluster and the cluster based on the similarity between the issue cluster and the cluster
Computer system, characterized in that. The method of claim 13,
The cluster generation unit,
Merging the issue cluster and the other issue cluster based on the similarity between the issue cluster and other issue clusters
Computer system, characterized in that. The method of claim 13,
The cluster generation unit,
Comparing the similarity between vectors of the clusters and merging clusters having a vector similarity meeting a predetermined cluster merging criterion into one cluster to create the issue cluster
Computer system, characterized in that. The method of claim 16,
The cluster generation unit,
When a document is added to the cluster, the vector of the cluster to which the document is added is calculated using the vector previously calculated for the cluster.
Computer system, characterized in that. The method of claim 16,
The cluster generation unit,
When another cluster is merged into the issue cluster, calculating the vector of the issue cluster in which the other clusters are merged using the vector previously calculated for the issue cluster.
Computer system, characterized in that. The method of claim 13,
The cluster generation unit,
In a parallel processing method, after comparing each cluster with the clusters created after the corresponding cluster, the most similar cluster is selected as a merge target,
In a serial processing method, the clusters selected as the merge target are sequentially merged
Computer system, characterized in that. The method of claim 13,
The cluster collection unit,
Collecting clusters clustered with similar documents based on the degree of similarity between documents
Computer system, characterized in that.

Images