KR101649884B1 - Method of maintaining recent frequent pattern based on damped window and approximation pruning in stream environment, recent frequent pattern maintenance server performing the same and storage medium storing the same - Google Patents

Abstract

The method of claim 1, wherein the latest frequent pattern maintenance method comprises: collecting a transaction based on a size of a first batch; mining at least one first frequent item if the transaction is collected by the size of the first batch, Collecting a transaction based on the size of the second batch, and, if the transaction is collected by the size of the second batch, at least one second frequent item is mapped And adding each of the at least one second frequent item to the tree as a node.

Description

TECHNICAL FIELD [0001] The present invention relates to a method for maintaining a frequent pattern based on attenuation windows and proximity pruning in a stream environment, a server for maintaining the latest frequent patterns, and a recording medium storing the updated frequent patterns. PATTERN MAINTENANCE SERVER PERFORMING THE SAME AND STORAGE MEDIUM STORING THE SAME}

The present invention relates to a technique for maintaining a frequent pattern, which is capable of collecting transactions based on a specific batch size, mining frequent items and inserting them into a tree, and deleting corresponding nodes for non-frequent items, The present invention relates to a method for maintaining the latest frequent patterns, a server for maintaining the latest frequent patterns for performing the same, and a recording medium for storing the updated frequent patterns.

In general, when transaction information is stored in a tree and maintained, and a request for frequent item extraction from a user is received, a frequent item is detected by mining a tree storing transaction information. Also, when a new transaction arrives, it reflects it in the tree, and performs mining operation on all transaction information whenever a request for frequent item extraction is received. That is, every time a new transaction arrives, the tree is updated by processing it, and a mining operation on the tree is required every time a request for frequent item extraction is received.

Korean Patent Registration No. 10-0899144 is directed to a method and apparatus for searching a maximum frequent item set from a data stream, wherein the occurrence frequencies of item sets and item sets included in already generated transactions are managed through a potential tree, It is a technique to find the most frequent item set by checking whether the corresponding item set is frequently visited while visiting the node.

Korean Patent No. 10-1105363 relates to a method for searching frequent item sets for a long transaction data stream, which divides a generated transaction into a plurality of divided transactions, and divides each of the plurality of divided transactions into a plurality of first hierarchical potentials It is a technique to generate a frequent item set by mining using a tree, to compress and merge the generated frequent item set, and to mining using a second-level dislocation tree.

Korean Patent No. 10-0899144 Korean Patent No. 10-1105363

One embodiment of the present invention is to provide a method for maintaining the latest frequent items using a single tree.

One embodiment of the present invention seeks to provide a method for maintaining the latest frequent patterns that can be used to mine frequent items after collecting transactions based on the size of a specific batch.

One embodiment of the present invention is to provide a method for maintaining the latest frequent patterns that can add frequent items to a tree as nodes and delete non-frequent items from the tree.

One embodiment of the present invention is to provide a method for keeping up with the latest frequent patterns that can search for frequent items stored in a tree to provide a set of the latest frequent items to a user.

Among the embodiments, the method for maintaining the latest frequent patterns comprises the steps of: (a) collecting a transaction based on a size of a first batch; (b) if the transaction is collected by the size of the first batch, (C) collecting a transaction based on a size of a second batch; and (d) determining whether the transaction is a second one of the second Mining at least one second frequent item and adding each of the at least one second frequent item as a node to the tree if collected by the size of the batch.

In one embodiment, if there is the same frequent item as the first frequent item stored in the tree among the second frequent items, the step (d) increases the frequency of the node including the first frequent item Step < / RTI >

In one embodiment, the step (d) may include updating the placement information of the node.

In one embodiment, if there is a frequent item that does not overlap with the second frequent item among the first frequent items stored in the tree, the step (d) may include calculating a frequency of the node storing the first frequent item .

In one embodiment, (e) calculating sequentially based on the number of transactions processed by the frequency of the node and the predetermined frequency reduction factor, the minimum frequency threshold, and the reliability rate determined by the user terminal, If the node is smaller than the nearest minimum frequency, the frequency of the node may be set to zero or the node may be deleted from the tree.

In one embodiment, if the corresponding node corresponds to a terminal node of the tree, the step (e) deletes the corresponding node, searches for the deleted node from the deleted node to the root node, .

In one embodiment, (f) if the user is requested to output the latest frequent item, searching for the frequent item stored in the tree may be further included.

In one embodiment, the step (f) searches for each node sequentially from the root node of the tree, and determines the number of transactions that the frequency of the node has processed so far and the frequency reduction factor and the minimum frequency And extracting a frequent item of the node included in the route from the corresponding node to the root node when the calculated value is larger than the calculated minimum frequency based on the threshold value.

In one embodiment, the tree may include a list of headers including links to connect nodes where frequent items and frequent items are stored.

In one embodiment, the node may include frequent items and frequency and placement information for the frequent items, and may include pointers to access parent and child nodes of the node.

Among the embodiments, the latest frequent pattern maintenance server comprises a transaction collector for collecting transactions based on the size of the first batch or the size of the second batch, at least one A tree generating unit for generating a tree by mining a first frequent item and each of the at least one first frequent item as a node and a tree generating unit for generating at least one second frequent item when the transaction is collected by the size of the second batch And a frequent item addition unit for mining and adding each of the at least one second frequent item as a node to the tree.

In one embodiment, the nodes of the tree are sequentially searched, and the number of transactions of the node is updated to the nearest minimum value calculated based on the frequency reduction factor, the minimum frequency threshold, And a frequent item management unit that sets the frequency of the corresponding node to 0 or deletes the corresponding node from the tree when the frequency is smaller than the frequency.

In one embodiment, when the user is requested to output the latest frequent item, the frequent item providing unit may search for and output the frequent item stored in the tree.

Among the embodiments, the recording medium on which the computer program relating to the method for maintaining the latest frequent patterns is recorded includes a function of collecting transactions based on the size of the first batch, a function of collecting the transactions based on the size of the first batch, A function of generating a tree by mining a first frequent item and each of the at least one first frequent item as a node, collecting a transaction based on the size of the second batch, and comparing the transaction with the size of the second batch And a function of mining at least one second frequent item when collected and adding each of the at least one second frequent item as a node to the tree.

The latest frequent pattern maintenance method according to an embodiment of the present invention can maintain and manage only the latest frequent items using one tree.

According to an embodiment of the present invention, the latest frequent pattern maintenance method can collect frequent items after collecting transactions based on the size of a specific batch.

In the method for maintaining the latest frequent patterns according to an embodiment of the present invention, a frequent item may be added to a tree as a node, and a non-frequent item may be deleted from the tree.

According to an embodiment of the present invention, the latest frequent pattern maintenance method can search a frequent item stored in a tree to provide a set of the latest frequent items to a user.

1 is a block diagram of a recent frequent pattern maintenance system according to an embodiment of the present invention.
2 is a block diagram of the latest frequent pattern maintenance server of FIG.
3 is a flowchart illustrating a method for maintaining the latest frequent patterns according to an embodiment of the present invention.
Figure 4 is a more specific flow diagram of a method for adding frequent items to a tree.
Figure 5 is a more specific flowchart of a method for managing non-frequent items in a tree.
Figure 6 is an example of a tree that maintains frequent items.

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, and also implemented in the form of a carrier wave (for example, transmission over the Internet) . In addition, the computer-readable recording medium may be distributed over network-connected computer systems so that computer readable codes can be stored and executed in a distributed manner.

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 of a recent frequent pattern maintenance system according to an embodiment of the present invention.

Referring to FIG. 1, the latest frequent pattern maintenance system 100 includes a user terminal 110, a latest frequent pattern maintenance server 120, and the like. Here, the user terminal 110 is connected to the latest frequent pattern maintenance server 120 via the network 130.

The user terminal 110 may correspond to a computing device connected to the latest frequent pattern maintenance server 120 and capable of transmitting and receiving data for performing the latest frequent pattern maintenance method. For example, the user terminal 110 may correspond to a desktop, a laptop, a smart phone, or a tablet personal computer. The user terminal 110 may send to the latest frequent pattern maintenance server 120 specific information necessary to set a specific reference value required to maintain the tree with the latest frequent items and may request the output of the latest frequent items, Information on the latest frequent items from the frequent pattern maintenance server 120 can be received.

The latest frequent pattern maintenance server 120 may correspond to a computing device that is connected to the user terminal 110 and can execute a specific task. In one embodiment, the latest frequent pattern maintenance server 120 can maintain and manage the latest frequent items in a tree form based on the specific information received from the user terminal 110, search for frequent items in the tree, (110).

2 is a block diagram of the latest frequent pattern maintenance server of FIG.

2, the latest frequent pattern maintenance server 120 includes a transaction collecting unit 210, a tree generating unit 220, a frequent item adding unit 230, a frequent item managing unit 240 and a control unit 250 do.

The transaction collecting unit 210 collects transactions by a corresponding size based on the size of a specific batch. In one embodiment, a transaction may correspond to one information document, for example, a transaction may correspond to a single webpage. In addition, the size of the batch may be pre-set by the user terminal 110.

The tree generating unit 220 generates a tree with each of the frequent items mined from the transaction collected by a specific batch size as a node. In one embodiment, referring to FIG. 6, the tree may include a list of headers, where the header list may include links to connect nodes where frequent items and frequent items are stored.

In one embodiment, each of the nodes in the tree may include frequent items and frequency and placement information for the frequent items, and may include pointers to access parent and child nodes of the node And links for connecting nodes with the same frequent items. Here, the frequency is a numerical value as to how frequently the frequent item appears in a specific batch, and the batch information corresponds to the information about how many batches the batch is mined. For example, if frequent item A is mined in a transaction collected based on the size of the third batch, and if the frequent item A appears 45 times in total in the third batch, the {frequency, placement information} Can be {45, 3}.

In one embodiment, the tree generating unit 220 may generate a tree based on the first frequent items mined from the collected transactions based on the size of the first batch corresponding to the first batch. That is, the tree generating unit 220 generates the tree only once in the latest frequent pattern maintenance process. In one embodiment, the tree generating unit 220 may generate a tree having each of the first frequent items as a node, and each node may include frequency and placement information for the frequent item.

In one embodiment, frequent items can be maintained and managed through other types of data structures other than trees, e.g., frequent items can be maintained and managed using a list-like data structure.

If the frequent item adding unit 230 generates a tree for the first frequent item through the tree generating unit 220, the frequent item adding unit 230 adds each of the second frequent items as a node to the tree generated through the tree generating unit 220 do. Here, the second frequent item corresponds to a frequent item mined from the collected transaction based on the size of the second batch corresponding to the second batch.

In one embodiment, if there are frequent items that are the same as the first frequent items stored in the tree among the second frequent items, the frequent item adding unit 230 may reduce the frequency of nodes including the first frequent item , The second frequent item is increased by the frequency and the placement information can be updated. For example, if {frequency, placement information} is stored as {45, 1} in a node of a specific first frequent item stored in the tree, and the second placement of a specific second frequent item that is the same as the specific first frequent item The frequency is reduced to {44 (= 45-1), 1} and {64 (= 44 + 20)} in {frequency and placement information} of the node of the first frequent item, 2}. ≪ / RTI > Here, the placement information may correspond to the batch currently being processed as the most recent batch in which a particular first frequent item appears.

In one embodiment, if there is a frequent item that does not overlap with the second frequent item among the first frequent items stored in the tree, the frequent item adding unit 230 adds the frequency of the node where the first frequent item is stored . For example, if the frequent item B corresponding to a specific first frequent item is stored in the tree and the {frequency and placement information} for the frequent item B corresponds to {15, 1}, and the frequent item B , The {frequency, placement information} for frequent item B can be modified to {14, 1}.

The frequent item management unit 240 sequentially searches the nodes of the tree, and if the frequency of the corresponding node is smaller than the nearest minimum frequency, the frequent item management unit 240 sets the frequency of the corresponding node to 0 or deletes the corresponding node from the tree. Here, the frequency of the node compared with the nearest minimum frequency corresponds to the re-calculated frequency multiplied by the difference between the currently stored batch information and the batch information last accessed to the node, Corresponds to a frequency reduction factor input from the user terminal 110. [

 In one embodiment, the nearest minimum frequency may be computed based on the number of transactions processed so far and the frequency reduction factor, the minimum frequency threshold, and the confidence rate predetermined by the user terminal 110. More specifically, the nearest minimum frequency can be calculated by [Equation 1] below.

[Equation 1]

μ: nearest minimum frequency

T _x : Number of transactions processed so far

λ: frequency reduction factor

minsup: Minimum frequency threshold (0 <minsup <= 1)

δ: Confidence rate

(=ε)는 오차값으로서, 즉, 근접 최소 빈도수는 최소 빈도수에서 오차를 뺌으로써 계산될 수 있다. 또한, 최소 빈도수(minsup)는 체르노프 바운드에 의하여 결정될 수 있다. <sub>Here, T x × λ (= Tx`</sub> ) is the number of the entire transaction, Tx` × minsup (= ξ) is the minimum frequency,

(= ε) is an error value, ie, the nearest minimum frequency can be calculated by subtracting the error from the minimum frequency. In addition, the minimum frequency (minsup) can be determined by chernopubund.

In an embodiment, if the node having a frequency smaller than the nearest minimum frequency corresponds to a terminal node of the tree, the frequent item management unit 240 deletes the corresponding node, searches for the root node from the deleted node, Node can be deleted. In addition, the frequent item management unit 240 may stop the process when a non-zero node appears from the deleted node to the root node.

In order to increase the memory efficiency of the tree, the frequent item management unit 240 may apply a chernoff bound to the item management unit 240 to increase the tree efficiency of the tree, Can be minimized.

In one embodiment, the latest frequent pattern maintenance server 120 may further include a frequent item provision unit, and when the frequent item provision unit receives an output of the latest frequent item from the user terminal 110, It is possible to search for a frequent item and provide it to the user terminal 110.

In one embodiment, the frequent item providing unit sequentially searches each node using the depth-first search method from the root node of the tree, and when the frequency of the node is greater than the minimum frequency, the frequent item providing unit includes the path from the corresponding node to the root node It is possible to extract a frequent item of a node. Here, the minimum frequency can be calculated based on the frequency reduction factor (?) And the minimum frequency threshold (minsup) predetermined by the user terminal 110, and is equal to the minimum frequency (?) Described in the formula 1 .

In one embodiment, when a particular task is being performed through the transaction collecting unit 210, the tree generating unit 220, the frequent item adding unit 230, or the frequent item managing unit 240, , And can extract frequent items after the work is completed.

The control unit 250 controls the operation of the transaction collecting unit 210, the tree generating unit 220, the frequent item adding unit 230 and the frequent item managing unit 240 and the flow of data do.

3 is a flowchart illustrating a method for maintaining the latest frequent patterns according to an embodiment of the present invention.

The transaction collecting unit 210 collects the transaction by a corresponding size based on the size of the first batch (step S301), and the tree generating unit 220 generates a tree by mining the first frequent item in the collected transaction Step S302). In one embodiment, the tree generating unit 220 may generate a tree having each of the first frequent items as a node, and each node may include frequency and placement information for the frequent item.

The transaction collecting unit 210 collects the transaction by a corresponding size based on the size of the second batch (step S302), and the frequent item adding unit 230 adds the second frequent item in the transaction collected based on the size of the second batch The item is added to the tree by mining (step S304). In one embodiment, the frequent item adding unit 230 creates a new node based on whether the second frequent item is the same as the first frequent item stored in the tree, and adds the frequent item to the tree or The frequency and arrangement information of the first frequent item stored in the tree can be updated.

Figure 4 is a more specific flow diagram of a method for adding frequent items to a tree.

The frequent item adding unit 230 mines a frequent item in a transaction collected by the size of a specific arrangement and arranges frequent items according to a predetermined reference by the user terminal 110 (step S401) Insert the frequent items.

The frequent item adding unit 230 determines whether the specific frequent item is the last frequent item to be inserted into the tree before inserting the specific frequent item into the tree (step S402).

In one embodiment, if the particular frequent item corresponds to the last frequent item, the frequent item adding unit 230 determines whether the frequent item exists in the child node of the node to which the previous frequent item is added (step S403). In one embodiment, if there is a frequent item in the child node, the frequency and the placement information of the child node are updated (step S404). Here, the frequency may be updated to a value obtained by adding the frequencies of the frequent items after the stored frequency is attenuated, and the placement information may be updated with the placement information for the frequent items. In another embodiment, if there is no frequent item in the child node, a node is newly created for the frequent item and inserted into the child node of the node to which the previous frequent item is added (step S405).

In one embodiment, if the particular frequent item does not correspond to the last frequent item, the frequent item adding unit 230 determines whether the frequent item exists in the child node of the node to which the previous frequent item is added (step S406) . Here, if a particular frequent item corresponds to a frequent item added to the first tree among the sorted frequent items, it can be checked whether or not the frequent item exists in the child node of the root node.

If there is a frequent item in the child node, the child node is set as the current node for the next frequent item (step S407), and the frequency and placement information of the corresponding child node are updated. Here, the update of the frequency and placement information can be updated in the same manner as that performed in step S404. More specifically, the frequent item addition unit 230 moves from the current node to the corresponding child node, and when the next frequent item is added, it is checked whether or not there is a frequent item in the child node of the corresponding child node based on the child node do. That is, the child node is set as the current node for the next frequent item. In another embodiment, if there is no frequent item in the child node, the same step S408 as in step S405 is performed.

Figure 5 is a more specific flowchart of a method for managing non-frequent items in a tree.

The frequent item management unit 240 searches each node of the tree by the depth first search method (step S501), and re-calculates the frequency of each node (step S502). In one embodiment, the frequency of each node is determined by multiplying the pre-stored frequency by a frequency reduction factor (lambda) previously set by the user terminal 110 by the difference between the current in-process deployment information and the last accessed deployment information Can be recalculated.

The frequent item management unit 240 searches each node and compares the recalculated frequency with a specific criterion to check whether the re-calculated frequency is smaller than a specific criterion (step S503). In one embodiment, a particular criterion may correspond to a nearest minimum frequency ([mu]).

In one embodiment, the frequent item management unit 240 determines whether the corresponding node corresponds to an internal node in a tree for a node whose re-calculated frequency is smaller than a specific criterion (step S504). Here, the internal node corresponds to a node other than the terminal node.

In one embodiment, if the node corresponds to an internal node, the frequent item management unit 240 may set the frequency of the corresponding node to 0 (step S505). In another embodiment, if the corresponding node does not correspond to an internal node, the frequent item management unit 240 may delete the corresponding node from the tree (step S506). In addition, the frequent item management unit 240 searches from the deleted node to the root node (step S507), and checks whether the frequency of each node is 0 (step S508). In one embodiment, if the frequency of a particular node corresponds to 0, the corresponding node is removed (step S509), and the process goes to step S507 to search for the next node.

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: The latest frequent pattern maintenance system
110: User terminal
120: Keeping up with the latest patterns
130: Network
210:
220: tree generating unit
230: Frequent item addition section
240: Frequent item management unit
250:

Claims (14)

Keeping up-to-date frequent patterns In a method for maintaining the latest frequent patterns performed in a server,
(a) collecting a transaction based on a size of a first batch;
(b) if the transaction is collected by the size of the first batch, mining at least one first frequent item and creating a tree with each of the at least one first frequent item as a node;
(c) collecting a transaction based on the size of the second batch; And
(d) if the transaction is collected by the size of the second batch, mining at least one second frequent item and adding each of the at least one second frequent item as a node to the tree, If there is a frequent item that does not overlap with the second frequent item among the first frequent items, decreasing the frequency of the node storing the first frequent item.
2. The method of claim 1, wherein step (d)
And decreasing the frequency of the node including the first frequent item if the frequent item is the same as the first frequent item stored in the tree among the second frequent items. How to keep up with the latest frequent patterns.
3. The method of claim 2, wherein step (d)
And updating the arrangement information of the node.
delete The method according to claim 1,
(e) searching sequentially the nodes of the tree, comparing the frequency of the node with the number of transactions processed up to now, and the proximity minimum frequency based on the frequency reduction factor, the minimum frequency threshold and the reliability rate determined by the user terminal And decreasing the frequency of the corresponding node to zero or deleting the node from the tree when the number of nodes is small.
6. The method of claim 5, wherein step (e)
If the corresponding node corresponds to a terminal node of the tree, deleting the corresponding node, searching for the deleted node to the root node, and deleting the node having the frequency of 0, Pattern preservation method.
The method according to claim 1,
(f) if the user is requested to output the latest frequent items, searching for the frequent items stored in the tree and providing the latest frequent items.
8. The method of claim 7, wherein step (f)
When each node is sequentially searched from the root node of the tree and the frequency of the corresponding node is larger than the calculated minimum frequency based on the number of transactions processed to date and the frequency reduction factor and the minimum frequency threshold preset by the user terminal And extracting a frequent item of a node included in a route from the corresponding node to the root node.
2. The method of claim 1,
And a header list including a link for connecting a node storing frequent items and frequent items.
2. The method of claim 1, wherein the node
And a pointer for accessing a parent node and a child node of the node, including frequency and placement information of a frequent item and a corresponding frequent item.
A transaction collector for collecting transactions based on the size of the first batch or the size of the second batch;
A tree generating unit for generating a tree by mining at least one first frequent item and each of the at least one first frequent item as a node when the transaction is collected by the size of the first batch; And
If the transaction is collected by the size of the second batch, mining at least one second frequent item and adding each of the at least one second frequent item to the tree as a node, And a frequent item addition unit for decreasing the frequency of the node in which the first frequent item is stored if there is a frequent item that does not overlap with the second frequent item among the frequent items.
12. The method of claim 11,
If the frequency of the corresponding node is sequentially searched for the nodes of the tree and is smaller than the proximity minimum frequency calculated on the basis of the number of transactions processed to date and the frequency reduction factor, the minimum frequency threshold and the reliability rate set by the user terminal Further comprising a frequent item management unit for setting the frequency of the corresponding node to 0 or deleting the corresponding node from the tree.
12. The method of claim 11,
Further comprising a frequent item providing unit for searching for and outputting a frequent item stored in the tree when the user is requested to output the latest frequent item.
Collecting the transaction based on the size of the first batch by the transaction collecting unit;
A function of generating a tree by mining at least one first frequent item and each of the at least one first frequent item as a node when the transaction is collected by the tree generating unit by the size of the first batch;
Collecting a transaction based on the size of the second batch by the transaction collecting unit; And
If the transaction is collected by the frequent item adding unit by the size of the second batch, mining at least one second frequent item and adding each of the at least one second frequent item to the tree as a node, And a function of reducing the frequency of the node in which the first frequent item is stored if there is a frequent item that does not overlap with the second frequent item among the first frequent items stored in the first frequent item storing unit A recording medium on which a computer program is recorded.

Images