KR101714435B1 - Method of mining a pattern related to input pattern, apparatus performing the same and storage medium storing a program performing the same - Google Patents

Abstract

The present invention relates to an associative pattern mining method and a server, wherein an associative pattern mining method includes: allocating a layout based on a set time interval; generating data based on a time range corresponding to the allocated layout; Generating a tree-based data structure, receiving a search time range and a search pattern, determining a search pattern for searching the search pattern based on the input search time range, Reconstructing a structure based on the input search pattern, and searching an association pattern in the corresponding layout based on the reconstructed tree-based data structure.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to an association pattern mining method, an association pattern mining method, an association pattern mining method, and a recording medium recording a program for performing the association pattern mining method.

The present invention relates to an associative pattern mining method, an associative pattern mining server, and a recording medium storing a program for executing the associative pattern mining method. More particularly, the present invention relates to an association A method for mining a pattern within a predetermined time range, a device, and a recording medium for recording a program for performing the method.

In a data stream environment, a method of mining a pattern based on a minimum threshold defined by a user can find patterns satisfying the condition in real time. However, this method has a problem that the mining result can not be limited only by the patterns associated with the specific pattern to be analyzed by the user.

In order to mine the patterns associated with a specific pattern, the minima must be changed by changing the minimum threshold, and the user must look at the mining results every time. This approach can take a lot of time to get the desired results.

Meanwhile, a method of mining a pattern based on a time interval or a time difference can not set a time range for a user to perform a mining operation. In addition to mining patterns based on a minimum threshold, There is a problem that it can not be done. That is, although the above schemes can mine the patterns satisfying the corresponding conditions based on the minimum threshold, it is difficult to efficiently detect the patterns associated with the inputted patterns without setting the minimum threshold.

Korean Patent Registration No. 10-1376444 relates to a pattern mining method for searching a tree in a top-down manner considering a weight in a data stream. The pattern mining apparatus includes a step of generating a first transaction in which data streams are sorted in descending weight order, Wherein the mining device inserts the first transaction into a Weighed Pattern-tree, the pattern mining device searches the WP-tree top-down, the pattern mining device uses the traversed path to perform a second transaction Generating a conditional database by using the second transaction as a database, and performing patterning work using the conditional database by the pattern mining apparatus.

Korean Patent Registration No. 10-0812378 is directed to a method for searching frequent itemsets using abbreviated disjoint trees for frequent item set search in a data stream environment, which is a continuously generated transaction data set. In the data stream environment, We define the effective data structure and use this data structure to search for necessary information. Korean Patent No. 10-0812378 defines a data structure as abbreviated dislocation tree structure and the abbreviated dislocation tree is compared with a dislocation tree structure applied to existing data mining to merge or separate nodes in the process of mining, The size of the tree can be flexibly adjusted dynamically by managing it in one node. This dynamic adjustment function dynamically merges and separates the nodes in the potential tree when the number of item sets that are likely to become a frequent item set is large due to a change in the data stream, That is, it is possible to maximize the accuracy of the frequent item set searched.

Korean Patent No. 10-1376444 (Mar. 13, 2013), pages 3 to 7 Korean Patent No. 10-0812378 (Mar. 4, 2008), pages 6 to 8

One embodiment of the present invention provides an association pattern mining method and a server capable of searching patterns within a set time range by reflecting data continuously generated in a stream environment into allocation allocated at regular time intervals in consideration of generation time I want to.

One embodiment of the present invention relates to an associative pattern mining method capable of searching for a time pattern by allocating a batch at a predetermined time interval using a time-sensitive sliding window and reflecting stream data generated within a time corresponding to each batch to the batch And a server.

In an embodiment of the present invention, when a time range is set and a pattern is input, an association pattern mining method capable of searching for and providing patterns related to the input pattern by each batch (time) using only stream data generated within the time range And apparatus.

Among the embodiments, the associative pattern mining method includes: allocating a layout based on a set time interval; generating a tree-based data structure based on data generated within a time range corresponding to the allocated layout; Determining a layout to search for the search pattern based on the input search time range, and reconstructing a tree-based data structure for the determined layout based on the input search pattern, And searching for a related intra-batch association pattern based on the reconstructed tree-based data structure.

In one embodiment, the step of allocating the layout includes receiving a time interval for the batch, comparing the time difference between the start time and the current time of the recently allocated batch with the input time interval, And if the difference is greater than the input time interval, reallocating the arrangement for the time after the input time interval elapses from the start time of the recently allocated arrangement.

In one embodiment, the tree-based data structure may have a structure that is grouped by placement.

In one embodiment, the tree-based data structure may include a table of nodes consisting of an item entry for storing item information and an item tree consisting of item nodes for storing information about the item.

In one embodiment, the header table may include an item name, a frequency, and a link indicating a node for the item in the potential tree.

In one embodiment, the item node may include an item name, a node link pointing to the same item created immediately before, a pointer to a parent node, and a list of child nodes.

In one embodiment, the step of reconstructing a tree-based data structure based on the input search pattern comprises reconstructing item nodes corresponding to items in the input pattern to be child nodes of item nodes corresponding to other items Step < / RTI >

In one embodiment, the step of reconstructing to be a child node of the item nodes corresponding to the other items comprises arranging the items in the input pattern according to the sort order of the header table in the tree data structure, Adjusting the order of items in the input pattern such that the items in the input pattern are downwardly shifted to items of another pattern while maintaining alignment; selecting items in the sorted input pattern in reverse order and searching for item nodes corresponding to the items selected in the potential tree; And changing the node order.

In one embodiment, when the item node corresponding to the selected item has a child node, the step of changing the order of the nodes while searching for the item node corresponding to the selected item may include: And switching the position of the item node corresponding to the selected item.

In one embodiment, the step of changing the node order while searching for the item node corresponding to the selected item may include the step of, when a sibling node having the same item exists for the child node after switching the position, And merging the sibling nodes.

In one embodiment, the step of changing the node order while searching for the item node corresponding to the selected item may include searching the next item node along the node link if the item node corresponding to the selected item does not have any child node . ≪ / RTI >

In one embodiment, the step of searching for an association pattern in the corresponding layout based on the reconstructed tree-based data structure comprises the steps of: accessing an entry in the header table corresponding to the last item in the sorted input pattern; Extracting all the paths from the item node to the root node connected through the corresponding entry and the link, and generating an association pattern by combining the remaining items in the extracted paths based on the input pattern .

In one embodiment, the step of searching for an association pattern in the corresponding arrangement based on the reconstructed tree-based data structure may include summing the frequencies in each arrangement within the search time range for each generated association pattern, And removing the associated pattern with a frequency of zero.

In embodiments, the associative pattern mining server includes a memory that stores computer-readable instructions and a processor that is electrically coupled to the memory and that performs the following steps over the stored instructions in response to a user request: Generating a tree-based data structure based on data generated within a time range corresponding to the allocated layout, receiving a search time range and a search pattern, Determining a layout to search for the search pattern based on the input search time range, reconstructing a tree-based data structure for the determined layout based on the input search pattern, and reconstructing the reconstructed tree- And executes a step of searching for an association pattern in the layout as a basis.

In one embodiment, the processor can reconstruct the tree-based data structure such that item nodes corresponding to items in the input pattern are child nodes of item nodes corresponding to other items.

In one embodiment, the processor arranges the items in the input pattern according to the sorting order of the header table in the tree data structure, and, within the header table, The order of nodes in the sorted input pattern may be changed in reverse order, and the order of the nodes may be changed while searching for the item node corresponding to the item selected in the potential tree.

In one embodiment, the processor accesses an entry in the header table corresponding to the last item in the sorted input pattern and extracts all paths from the item node to the root node linked through the link with the corresponding entry , And may generate an association pattern by combining the remaining items existing in the extracted paths based on the input pattern.

In one embodiment, the processor may sum the frequencies in each batch within the search time range for each generated association pattern, and remove the association pattern with the combined frequency of zero.

Among the embodiments, the recording medium storing the computer program relating to the associative pattern mining method includes a function of allocating a layout on the basis of a set time interval, a function of allocating a tree based on the data generated within the time range corresponding to the allocated layout, A function for generating a data structure, a function for receiving a search time range and a search pattern, a layout for searching the search pattern based on the input search time range, and a tree-based data structure for the determined layout, Based on the reconstructed tree-based data structure, and a function of searching for an association pattern in the corresponding layout based on the reconstructed tree-based data structure.

The associative pattern mining method and the server according to the embodiment of the present invention can search patterns within a set time range by reflecting the data continuously generated in the stream environment to the allocation allocated at regular time intervals considering the generation time .

The association pattern mining method and the server according to an embodiment of the present invention allocate layouts at predetermined time intervals using a time-sensitive sliding window, and reflect the generated stream data in a corresponding time according to each layout, Pattern search is possible.

The association pattern mining method and the server according to an embodiment of the present invention search patterns related to an input pattern by batch (time) using only stream data generated within a corresponding time range by setting a time range and inputting a pattern .

1 is a block diagram illustrating an associative pattern mining system according to an embodiment of the present invention.
2 is a block diagram illustrating the association pattern mining server of FIG.
FIG. 3 is a flowchart illustrating an associative pattern mining method executed in the associative pattern mining server in FIG. 2; FIG.
4 is a diagram for explaining a tree-based data structure.
5 is a flowchart illustrating a process of constructing a tree-based data structure.
6 is a diagram illustrating a process of reconstructing a tree-based data structure.
7 is a flowchart for explaining a process of searching an association pattern associated with an input pattern within a set time range.

The description of the present invention is merely an example for structural or functional explanation, and the scope of the present invention should not be construed as being limited by the embodiments described in the text. That is, the embodiments are to be construed as being variously embodied and having various forms, so that the scope of the present invention should be understood to include equivalents capable of realizing technical ideas. Also, the purpose or effect of the present invention should not be construed as limiting the scope of the present invention, since it does not mean that a specific embodiment should include all or only such effect.

Meanwhile, the meaning of the terms described in the present application should be understood as follows.

The terms "first "," second ", and the like are intended to distinguish one element from another, and the scope of the right should not be limited by these terms. For example, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component.

It is to be understood that when an element is referred to as being "connected" to another element, it may be directly connected to the other element, but there may be other elements in between. On the other hand, when an element is referred to as being "directly connected" to another element, it should be understood that there are no other elements in between. On the other hand, other expressions that describe the relationship between components, such as "between" and "between" or "neighboring to" and "directly adjacent to" should be interpreted as well.

It is to be understood that the singular " include " or "have" are to be construed as including the stated feature, number, step, operation, It is to be understood that the combination is intended to specify that it does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

In each step, the identification code (e.g., a, b, c, etc.) is used for convenience of explanation, the identification code does not describe the order of each step, Unless otherwise stated, it may occur differently from the stated order. That is, each step may occur in the same order as described, may be performed substantially concurrently, or may be performed in reverse order.

The present invention can be embodied as computer-readable code on a computer-readable recording medium, and the computer-readable recording medium includes any type of recording device that stores data that can be read by a computer system . Examples of the computer-readable recording medium include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like. , Computer-readable code in a distributed fashion can be stored and executed.

All terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined. Commonly used predefined terms should be interpreted to be consistent with the meanings in the context of the related art and can not be interpreted as having ideal or overly formal meaning unless explicitly defined in the present application.

1 is a block diagram illustrating an associative pattern mining system according to an embodiment of the present invention.

Referring to FIG. 1, an associative pattern mining system 100 includes a user terminal 110 and an associative pattern mining server 120.

The user terminal 110 may access the association pattern mining server 120 under the control of the user and request association pattern mining. The user terminal 110 displays the progress information or the result information provided from the association pattern mining server 120 on the screen. The user terminal 110 may correspond to a mobile terminal, a tablet PC, a laptop PC, or a desktop PC.

The association pattern mining server 120 mines an association pattern associated with an input pattern according to a request of the user terminal 110 in a stream environment in which data is continuously generated.

The association pattern mining server 120 processes data generated continuously in a stream environment on a time-sensitive sliding window basis, and the window is composed of a number of batches defined by a user. In addition, each batch is allocated at regular time intervals, and stream data generated at a time corresponding to the corresponding time interval is inserted into the batch.

The association pattern mining server 120 inserts stream data into the layout in consideration of generation time, and generates a tree-based data structure based on the data. The associative pattern mining server 120 can minify the pattern within a set time range based on the generated tree-based data structure. That is, the association pattern mining server 120 allocates the layouts at predetermined time intervals using the time-sensitive sliding window, and reflects the generated stream data within the time corresponding to each layout to the corresponding layouts to perform the time-based pattern mining .

For example, the association pattern mining server 120 allocates a layout based on a set time interval, and generates a tree-based data structure based on data generated within a time range corresponding to the allocated layout. When a search time range and a search pattern are input, the association pattern mining server 120 determines a layout to search for a search pattern based on the input search time range, and searches for a tree- Reconstruct based on pattern. The association pattern mining server 120 searches for an association pattern in the corresponding layout based on the reconstructed tree-based data structure. The association pattern mining server 120 may provide the user terminal 110 with the mining progress information and the result information.

2 is a block diagram illustrating the association pattern mining server of FIG.

Referring to FIG. 2, the associative pattern mining server 120 includes a processor (or CPU) 210 that communicates with other components via a bus 220. Processor 210 controls the operation of other components and performs associative pattern mining with other components. The processor 210 is also electrically connected to the memory 230 and may perform associative pattern mining through instructions stored in the memory 230 at the request of the user.

The associative pattern mining server 120 includes a memory 230 and a storage device 240. The memory 230 includes a read only memory (ROM) 232 and a random access memory (RAM) . Here, the memory 230 corresponds to a temporary or permanent computer-readable recording medium, and the storage device 240 corresponds to a permanent computer-readable recording medium. At least one of the memory 230 and the storage device 240 stores a computer program including instructions for associating pattern mining.

The association pattern mining server 120 includes a network interface 270 for communicating with the network 280. The network interface 270 establishes an environment in which information, data, and signals between the association pattern mining server 120 and the network 280 can be transmitted. Referring to FIG. 1, the association pattern mining server 120 may be connected to the user terminal 110 through a network 280.

The associative pattern mining server 120 may include a user interface input device 250 (e.g., a mouse, a trackball, a touchpad, a graphic tablet, a scanner, a barcode scanner for product barcode scanning, a touch screen, a keyboard, a pointing device, etc.) Information can be input from the user through the Internet. The user interface input device 250 includes all the mechanisms by which information (e.g., transactions) can be entered into the association pattern mining server 120 or the network.

The association pattern mining server 120 may output information through the user interface output device 260. [ The user interface output device 260 may include, but is not limited to, a visual output device such as a display screen. The user interface output device 260 may include any mechanism capable of outputting information to a user and may be combined with an output device such as a video output device or a speaker.

In one embodiment, the display screen may display information received from the associative pattern mining server 120 and receive input from an operator. That is, the display screen may be implemented as a user interface input device 250 and a user interface output device 260.

FIG. 3 is a flowchart illustrating an associative pattern mining method executed in the associative pattern mining server in FIG. 2; FIG.

Referring to FIG. 3, the association pattern mining server 120 allocates a batch (314) for each set time interval (T _T ) 312 based on the start time in the stream environment. In one embodiment, the association pattern mining server 120 may receive the time interval T _T through the user terminal 110.

When the batch is assigned, the association pattern mining server 120 continuously reflects the stream data generated within the time range corresponding to the allocated batch to the batch in the tree-based data structure 324 (125). In one embodiment, the tree-based data structure 324 may have a structure that is grouped by placement.

The user terminal 110 may receive the search time range and the search pattern and transmit the search time range and the search pattern to the association pattern mining server 120 under the control of the user. If the search time range and the search pattern are input 330, the association pattern mining server 120 first searches the tree-based data structure 324 for searching for the association patterns based on the input search pattern (332).

Next, the associative pattern mining server 120 calculates the batches corresponding to the set search time range. That is, the association pattern mining server 120 determines a search pattern for searching for a search pattern based on the input search time range.

The associative pattern mining server 120 searches for (336) at least one association pattern associated with the input pattern at the corresponding batches in the tree-based data structure and provides batch-specific search results (338). The user terminal 110 may display the provided search result on a screen and provide the search result to a user.

Hereinafter, the association pattern mining method will be described in detail with reference to FIG.

The user terminal 110 may receive the time interval T _T of the batch by the user's control and transmit the time interval T _T to the association pattern mining server 120. When the time interval T _T of the batch is input 312, the associative pattern mining server 120 allocates 314 the batch for the current time. Next, in a stream environment in which data is continuously generated, the association pattern mining server 120 stores the start time and the current time of the most recently allocated batch in the variables T _P and T _C , respectively (316).

The associating pattern mining server 120 calculates the time difference between the start time T _P and the current time T _C of the continuously allocated batch and stores the calculated time difference and the set time interval T _T (318). If the start time (T _P) and the current time (T _C), the time interval (T _T) is set from the time interval (T _T) not smaller than the (larger) from the previous batch time (T _P), the time difference is set between by The batch for the last time is reassigned (314).

If the time difference between the start time (T _P ) and the current time (T _C ) is less than the set time interval (T _T ), the association pattern mining server 120 determines whether new data has been received (320). If no new data is received, the association pattern mining server 120 returns to step 318 of comparing the time difference between the start time (T _P ) and the current time (T _C ) with the set time interval (T _T ) .

When the new stream data is received, the association pattern mining server 120 inserts the received data into the time-sensitive sliding window based tree structure 324 (322). The inserted stream data is reflected in the current placement information of the tree data structure 324. Through this process, the associative pattern mining server 120 can generate a tree-based data structure based on data generated within a time range corresponding to the allocated layout.

4 is a diagram for explaining a tree-based data structure.

Referring to FIG. 4, a time-sensitive sliding window-based tree data structure 410 includes a header table 410 and a dislocation tree 420. The tree data structure of FIG. 4 is an example in which two batches are allocated.

The header table 410 is composed of entries for storing information on items, and each entry is composed of an item name, a frequency (Support), and a link. The link points to the node for the last item created in the dislocation tree 420.

The dislocation tree 420 consists of at least one item node. Each item node includes an item name, a node link indicating a node for the same item generated immediately before, a pointer indicating a parent node, and a child node list for managing a set of child nodes. In addition, each item node includes a batch frequency array 430 for storing batch frequency information to be based on a time-sensitive sliding window.

5 is a flowchart illustrating a process of constructing a tree-based data structure.

5, if a layout for the current time is first allocated 512, the association pattern mining server 120 searches 514 for each item node in the tree data structure 324, A new element is added 516 to the end of the array 430 and the value is set to 0 (518). Through this process, we can construct a tree data structure to which a new batch is assigned.

When the new stream data is received 530, the associative pattern mining server 120 arranges the items in the data according to the sort order of the header table 410 in the tree data structure 324 (532) Each item is inserted in turn (534). The association pattern mining server 120 increments the last value of the arrangement frequency array for the item node corresponding to the item to be inserted by 1 (536).

Referring again to FIG. 3, the user terminal 110 receives the search time range and the search pattern and transmits the search time range and the search pattern to the association pattern mining server 120 under the control of the user. If the search time range and the search pattern are input 330, the association pattern mining server 120 reconstructs the tree-based data structure so that the item nodes for the items in the input pattern become the child nodes of the item nodes for the other items (332).

6 is a diagram illustrating a process of reconstructing a tree-based data structure.

Referring to FIG. 6, the association pattern mining server 120 first aligns the items of the input pattern 612 according to the sort order of the header table 410 in the tree data structure 324 (614).

The association pattern mining server 120 adjusts the order of the items in the input pattern in the header table 410 so that the items in the input pattern descend below other items while maintaining the sorting order. For example, assuming that the sorted input pattern for the header table 410 of FIG. 4 with the sort order {D, C, A, B, E} is {A, B} C, E, A, and B} in the header table 410. The header table 410 includes a header table 410,

The associating pattern mining server 120 then selects 610 each item in the sorted pattern in the header table 410 and accesses 620 an entry for the selected item in the header table 410.

The association pattern mining server 120 accesses each item node for the corresponding item through the link of the accessed entry and the node link of the item node connected to the link (622). The association pattern mining server 120 determines whether the accessed item node includes a child node (624).

If the node does not have a child node, the association pattern mining server 120 accesses the next item node along the node link (622). If the node has a child node, the associative pattern mining server 120 selects an item node having the highest sorting order among the child nodes to exchange (switch) the positions of the two nodes (626). That is, the association pattern mining server 120 exchanges the item node having the highest sorting order among the child nodes and the item node corresponding to the selected item.

After exchanging the positions, if there are siblings having the same item for the child node, the association pattern mining server 120 merges the two item nodes. For example, if item node C: {1, 2} is exchanged with parent node D: {7, 5} in the potential tree 420 of FIG. 4, 2} becomes an item node C: {2, 3} and a sibling node, the association pattern mining server 120 merges the two nodes. The association pattern mining server 120 recursively performs the process for the lower nodes (628).

When the above process is completed for all the items in the sorted input pattern, the process of reconstructing the tree-based data structure is completed.

Referring again to FIG. 3, the associative pattern mining server 120 computes (510) arrangements within the time range to search for the association patterns for input patterns within the time range input from the reconstructed tree data structure 334, . That is, the associative pattern mining server 120 determines a layout to search for an input pattern based on the input search time range.

The associative pattern mining server 120 searches 336 for the determined intra-batch intra-batch related patterns and provides the detected per-batch related pattern results 338. [

7 is a flowchart for explaining a process of searching an association pattern associated with an input pattern within a set time range.

Hereinafter, a process of searching for an association pattern will be described in detail with reference to FIG.

When the reconstruction process of the tree data structure is completed, the association pattern mining server 120 first calculates 714 the layouts within the search time range input. The association pattern mining server 120 accesses an entry corresponding to the last item in the sorted input pattern in the entry in the header table 410 (716).

The association pattern mining server 120 accesses the item node connected to the link of the corresponding entry (718), and extracts all the routes from the corresponding item node to the root node (720). At this time, each item node in the extracted path has frequency values for the corresponding arrangements in the arrangement frequency array 430.

For example, assuming that only the second batch exists within the search time range input in FIG. 4, the item node B: the item node in the path {B, A, C, D} Have arrangement frequency arrangement 230 of {1}, {2}, {2}, and {5}, respectively. In the above example, if there are two arrangements in the input search time range, the item nodes in the extracted route are {0,1}, {0,2}, {1,2}, { .

After extracting the path, the associative pattern mining server 120 moves to the next item node along the node link (718), and repeats the above process until there is no connected item node (722).

When the extraction of all the paths is completed, the association pattern mining server 120 generates an association pattern by combining the remaining items in the extracted paths based on the input pattern (724). For example, when the input pattern is {B, A} for the extracted path {B, A, C, D} of the above example, the result of combining the input pattern and the remaining items {C, D} , A, C}, {B, A, D}, {B, A, C, D}.

The frequency of each associated association pattern is determined by the smallest item frequency calculated in each determined layout. For example, each item in the combined pattern {B, A, C} in FIG. 4 has a frequency of {0, 1}, {0, 2}, {1, 2} The frequency is 0 and 1. Thus, {B, A, C} is a pattern that did not appear in the first batch, but a pattern with a frequency of one in the second batch.

Associated pattern mining server 120 consolidates the frequencies in a manner that adds the frequencies in each batch within the search time range for each association pattern (726). For example, if the frequency for the first batch and the frequency for the second batch are 1 and 2, respectively, for the pattern {B, A, C} among the patterns associated with the input pattern {B, A} The frequency is 3.

In step 728, the association pattern mining server 120 provides a combination pattern having a frequency of 1 or more.

Through the above process, the associative pattern mining server 120 searches and integrates association patterns of patterns inputted on the basis of a time-sensitive sliding window in a stream environment where data is continuously generated, have.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit and scope of the present invention as defined by the following claims It can be understood that

100: Associative Pattern Mining System
110: User terminal
120: Associative Pattern Mining Server

Claims (19)

1. An association pattern mining method performed in an associative pattern mining server,
Assigning a layout based on a set time interval;
Generating a tree-based data structure based on data generated within a time range corresponding to the allocated layout;
Receiving a search time range and a search pattern;
Determining a layout to search for the search pattern based on the input search time range, and reconstructing a tree-based data structure for the determined layout based on the input search pattern; And
And searching for an association pattern in the corresponding layout based on the reconstructed tree-based data structure. 2. The method of claim 1, wherein assigning
Receiving a time interval for the batch;
Comparing a time difference between a start time and a current time of a recently allocated batch with the input time interval; And
And if the time difference is greater than the input time interval, reassigning the arrangement for the time after the input time interval elapses from the start time of the recently allocated arrangement. Way. The method of claim 1, wherein the tree-based data structure
Wherein each of the plurality of patterns has a structure divided according to a layout. 4. The method of claim 3, wherein the tree-based data structure
A header table composed of item entries for storing item information, and an item tree storing item information about the items. 5. The method of claim 4, wherein the header table
An item name, a frequency, and a link indicating a node for the item in the dislocation tree. 5. The method of claim 4, wherein the item node
An item name, a node link indicating a node for the same item generated immediately before, a pointer indicating a parent node, and a list of child nodes. 5. The method of claim 4, wherein reconstructing the tree-based data structure based on the input search pattern comprises:
And reconfiguring item nodes corresponding to items in the input search pattern to be child nodes of item nodes corresponding to other items. 8. The method of claim 7, wherein reconstructing the item nodes to be child nodes of the item nodes corresponding to the other items
Arranging items in the input search pattern according to a sort order of a header table in the tree-based data structure;
Arranging the items in the input search pattern in the header table so that the items in the input search pattern descend under items of another pattern while maintaining alignment; And
Selecting items in the sorted search pattern in reverse order and searching for item nodes corresponding to the selected item in the potential tree and changing the node order. 9. The method of claim 8, wherein changing the node order while searching for an item node corresponding to the selected item
If the item node corresponding to the selected item has a child node, switching the item node having the fastest sorting order among the child nodes and the position of the item node corresponding to the selected item, Mining method. 10. The method of claim 9, wherein changing the node order while searching for an item node corresponding to the selected item
Further comprising merging the child node and the sibling node if there is a sibling node having the same item for the child node after switching the location of the sibling node. 9. The method of claim 8, wherein changing the node order while searching for an item node corresponding to the selected item
And if the item node corresponding to the selected item does not have a child node, searching for the next item node along the node link. 9. The method of claim 8, wherein the step of searching for an association pattern in the corresponding layout based on the reconstructed tree-
Accessing an entry in the header table corresponding to the last item in the aligned search pattern;
Extracting all paths from the item node to the root node linked through the link with the corresponding entry; And
And generating an association pattern by combining the remaining items in the extracted paths based on the input search pattern. 13. The method of claim 12, wherein the step of searching for an association pattern in the corresponding layout based on the reconstructed tree-
Summing the frequencies in each arrangement within the search time range for each generated association pattern; And
And removing the association pattern in which the combined frequency is zero. A memory for storing computer readable instructions; And
And a processor electrically connected to the memory and executing the following steps through the stored instructions according to a user's request,
The processor
Assigning a layout based on a set time interval;
Generating a tree-based data structure based on data generated within a time range corresponding to the allocated layout;
Receiving a search time range and a search pattern;
Determining a layout to search for the search pattern based on the input search time range, and reconstructing a tree-based data structure for the determined layout based on the input search pattern; And
And searching for an association pattern in the corresponding batch based on the reconstructed tree-based data structure. 15. The apparatus of claim 14, wherein the processor
Wherein the tree-based data structure is reconfigured so that item nodes corresponding to items in the input search pattern become child nodes of item nodes corresponding to other items. 16. The system of claim 15, wherein the processor
Arranging the items in the input search pattern according to the sort order of the header table in the tree-based data structure, and arranging the items in the input search pattern in the header table Descending order of the items in the ordered search pattern, and selecting the items in the ordered search pattern in reverse order and searching for the item node corresponding to the selected item in the potential tree, thereby changing the node order. 17. The system of claim 16, wherein the processor
Accesses an entry corresponding to the last item in the sorted search pattern in an entry in the header table, extracts all routes from the item node connected to the corresponding entry through the link to the root node, To generate an association pattern by combining the extracted items with remaining items existing in the extracted paths. 18. The apparatus of claim 17, wherein the processor
And summing the frequency numbers in each batch within the search time range for each of the generated association patterns, and removing the association pattern in which the combined frequency is zero. A computer-readable recording medium on which a program for implementing an associative pattern mining method is recorded,
Assigning a layout based on a set time interval;
Generating a tree-based data structure based on data generated within a time range corresponding to the allocated layout;
Receiving a search time range and a search pattern;
Determining a layout to search for the search pattern based on the input search time range, and reconstructing a tree-based data structure for the determined layout based on the input search pattern; And
And searching for an association pattern in the corresponding layout based on the reconstructed tree-based data structure.

Images