JP3466152B2 - Item attribute filtering method, item attribute filtering device, and recording medium - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique (filtering) for selecting various items (users, traders, information, CDs, books, etc.), and more specifically, based on attributes assigned to items. The present invention relates to an item attribute filtering method and device, and a recording medium having the program recorded therein.

[0002]

2. Description of the Related Art Conventionally, the process of classifying items based on the attributes given to the items has been widely used in recommended introduction services using the Internet, direct mail of general sellers, etc. Uses a method (specified attribute classification) of classifying items based on attribute conditions. On the other hand, there is a method in which a certain item is used as a reference and another item group is selected according to the attribute of the item (user attribute classification). This is used to find similar items.

[0003]

However, if both of these are to be supported, the two classifications will have different attribute processing settings due to different classification criteria. In that case, since the process occurs twice, the system becomes bloated, complicated, and the speed is inevitably reduced. This is especially noticeable in services that use both of these together.

In order to solve the above problems, the present invention aims to provide an item attribute filtering method and apparatus for unifying and processing designated attribute classifications and user attribute classifications, and a recording medium recording the program thereof. To do.

[0005]

[Means for Solving the Problem] First, as a precondition,
Items to be filtered (target items)
Each of the sets has attribute data of n bits. The n bits are divided into k (k is a positive integer) attribute categories (attribute value sets), each attribute category contains one or more bits, and each of the attribute category bits is unique within the category. Has been given meaning. For example, in FIG.
As shown in (a), when n = 16, 3 of A, B, and C
When divided into three (k = 3), in the section A, meanings A1, A2, A3, A4, A5 are given from the right. A plurality of bits may be set in A1 to A5, or all bits may be 0. The same applies to divisions B and C.

Now, it is assumed that there is a main item for which filtering processing is performed on such a target item. This main item also has n-bit attribute data.
Now, (a) when filtering by some of the attribute classifications of the main item, (b) when filtering by specifying the attribute classifications independently of the main item, and
(C) Consider both.

When the number of attribute categories that can be filtered is c, an n-bit attribute filter ... C main item merge determination flag ... C bits are prepared as filter conditions. Each bit of the main item merge determination flag has a one-to-one correspondence with each of the c attribute filters and is 1 (true).
Means valid and 0 (false) means invalid. The filtering process is performed in the following procedure. (1) A bit (flag bit) of the main item merge determination flag associated with each of the c attribute filters
I see. (2) If the flag bit is true , the logical product of the attribute filter and the main item attribute is taken, and the result is replaced with the original attribute filter (specific item attribute value set filter condition generation). If false, leave it alone. (3) If there are remaining attribute filters with all 0s, remove them. (4) The attribute filters remaining as a result are sequentially compared with the attribute data of the target item one by one. Therefore, if true (if all are not 0), the next attribute filter similarly compares. If all the attribute filters are true, the target item is true.

Specifically, the above-mentioned (a), (b), (c)
The procedure for realizing is shown below. (A) When filtering with some of the attribute classifications of the main item It is determined whether the attribute classification to be filtered matches the main item. Specifically, the main item merge determination flag is set to all 1, and the c attribute filters are set to all 1 for each attribute classification.
For example, in the case of the example in FIG.
The attribute filter and the merge determination flag are set as shown in FIG. As a result, if the attribute data of the main item is FIG. 3C, the attribute filter after replacement (specific item attribute value set filter condition) is as shown in FIG. 3D. The attribute filters are sequentially compared with the target item attribute data one by one.

(B) When the attribute category is designated independently from the main item for filtering, all the main item merge determination flags are set to 0. Compare with target item attribute data as it is with the given attribute filter.

(C) In the case of the combination of both (a) and (b) Only 1 is set in the main item merge determination flag indicating the bit to be filtered by the main item.
The attribute filter whose merge flag is 1 is (a)
Similarly to the above, the attribute data of the main item is multiplied and the attribute filter is replaced. The attribute filter whose merge flag is 0 is the same as that in (b), that is, it is left as it is.
The result is again used as an attribute filter (specific item attribute set filter condition) and sequentially compared with the attribute data of the target item.

As described above, the filter processing according to the present invention satisfies the cases (a), (b), and (c) described above, and c filters can be shared.

[0012]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described in detail below with reference to the drawings. FIG. 1 is a system configuration diagram of an embodiment of the present invention. In the figure, 100 is a user terminal, 200 is an item attribute filtering apparatus according to the present invention, and both are generally connected via the Internet or other networks.

The user terminal 100 is a terminal such as a personal computer, a digital interactive television, an information electrician or the like. The user sends a search condition (introduction request) and a search condition to the item attribute filtering apparatus 200 by the user terminal 100, and the item attribute filtering apparatus 200 responds to the user terminal 100 with the item of the search result. Note that the item attribute filtering apparatus 200 periodically retrieves the search result by direct email or the like, for example, from the user terminal 1.
It is also possible to send it to 00.

Item attribute filtering device 200
Are roughly divided into a filter value management unit 210 as a filter management database, an item information management unit 220 as an item management database, and a filtering execution unit 230 as a data processing device main body . The filter value management unit 210 manages the bit value of each attribute filter and the main item merge determination flag. Item information management unit 22
0 manages the item ID and item attribute data of each item. Of course, the item information management unit 220 also manages personal information of each user of each user terminal 100 as an item. Filtering execution unit 230
Extracts the attribute filter and the main item merge determination flag from the filter value management unit 210 based on the search condition, and the item attribute data of the item that is the main item and the target item of the item that is the main item from the item information management unit 220. Attribute data is taken out and filtering processing is executed.

FIG. 2 shows an embodiment of the processing flow in the filtering execution unit 230. Filtering execution unit 23
0 indicates that c attribute filters (each n bits) from the filter value management unit 210 based on the given search condition.
And the main item merge determination flag (c bit) are read (steps 11 and 12). That is, the filter condition is acquired. Then, it is determined whether the bit values of the main item merge determination flag are all 0s (step 13), and all 0s are set.
If not, the attribute data of the subject item to be filtered is read from the item information management unit 220 based on the given search condition (step 14).

Next, the filtering execution unit 230
In each attribute filter, for each attribute filter in which the corresponding bit of the main item merge determination flag is 1, the logical product of each bit is taken with the main item attribute data, and the result is replaced with the original attribute filter (step 15). . That is, the specific item attribute value set filter condition is generated. The attribute filter whose main item merge flag is 0 is left as it is. When it is determined in step 13 that the subject item merge determination flag bit value is all 0, each attribute filter extracted in step 11 is left as it is. If any of the remaining attribute filters have bit values of all 0, they are removed.

Next, the filtering execution unit 230 sequentially reads out the attribute data of the item to be filtered from the item information management unit 220 based on the search condition, takes the logical product of each attribute filter bit by bit, and the result is If not all 0, the same comparison as the next attribute filter is repeated (steps 16 to 19). As a result,
If it is determined that all the attribute filters are not all 0, the comparison result of the target item is set to be true (step 20). Further, if all 0 are judged by any of the attribute filters, the comparison result of the target item is set to false at that time (step 21). This is repeated for each target item (step 22).

The computer program for realizing the processing flow shown in FIG. 2 may be recorded in advance in a computer-readable recording medium, for example, a floppy disk, a CD-ROM, a memory card or the like and provided. It is possible.

A specific processing example will be described below. It is assumed that a user of a certain service inputs attribute data when registering the service, and this is managed by the item information management unit 220. The attribute data is 16 bits (n = 16),
Attribute category for the user to set There are three, used as follows:
It is used . Category A: Gender: 2 bits (female, male) Category B: Age: 4 bits (40s or older, 30s, 2)
0s, 10s and under) Category C: Hobbies: 5 bits (reading, fishing, sports,
Games, travel) Here, for example, 2 bits of category A (sex) (female, male)
2) of the attribute classification from left to female and male
It means giving meaning. The same applies to divisions B and C.
is there. As the remaining divisions D, for example, those which can be independently set by the service administrator are prepared. Category D: For management 5 bits (dealer member, deleted member, suspended member, reserve, reserve) Therefore, the number of attribute categories (attribute value sets) is 4, and bit allocation is as shown in FIG. .

(1) Service for finding a person who has the same interest as user A In order to find a similar friend, another user (target item) having the same attribute as user A (main item)
Find out. The number of filters c is 3. For example, if the user has an attribute of "20s, men, fishing or sports", the attribute data (attribute data of the main item) is as shown in Fig. 5A. Further, the attribute filter and the main item merge determination flag are set as shown in FIG. The reason for specifying the bits from the lower bit is that they can be applied as they are when the number of filters is expanded in the future.

Since the merge determination flag is all 1,
All the attribute filters are applied to the attribute data of the user A (attribute data of the main item). As a result, FIG.
The original attribute filter in (b) is replaced as shown in FIG. 5 (c).

For example, FIG. 5D shows a comparison with the attribute data in which the attribute of a certain user X (target item) is "20s, male, sports". Since this comparison result is all true, this user X is extracted and introduced to the user A. Thus, the user X and the user A do not have the same attribute data, but can be extracted if there is at least one common part in the attribute classification.

(2) Service for Extracting Users with Certain Attributes Consider a service for creating a list of users (target items) who are likely to be interested in order to send a direct electronic mail to introduce a certain product. Assuming that the product is cosmetics, a "woman in her twenties or thirties" is found.

In this example, since the specific user A (main item) is not assumed, the merge determination flag becomes all 0s. The attribute filter and the main item merge determination flag are as shown in FIG. Merge flag is all 0
Therefore, all the attribute filters are used as they are, as shown in FIG.

For example, FIG. 6C shows a comparison with the attribute data in which the attribute of a certain user Y (target item) is “30s, female, travel”. As a result, this user Y is extracted. If you set the filter like this, 2
Both 0s and 30s can be extracted.

(3) Dealer Matching Service Consider a service that introduces a seller to each user as an agent for selling products. It is assumed that the distributor has the distributor member attribute on the system.

The member of the trader can set the desired user attribute. For example, if a high-end fishing rod company Z, "20
For generations and up, both men and women are targeted for users of "fishing or sports". The attribute data is as shown in FIG.
All target attributes are set to 1. on the other hand,
For example, if the user B has an attribute of "40s, male, fishing", the attribute data will be as shown in Fig. 7B.

Therefore, the attribute filter and the main item merge determination flag are set as shown in FIG. 7 (c). This means to extract dealers whose ages and hobbies are the same. The filter 1 is for instructing the seller, and the merge flag is 0. This is because an ordinary user is not a trader, and when the trader attribute flags are merged, all 0s are obtained, and the function as a filter cannot be achieved.

Since the filters 2 and 3 have the merge flag 1, they are applied to the attribute data of the user B. As a result, the obtained attribute filter is as shown in FIG. Figure 7
When the attribute filter of (d) is compared with the attribute data of the trader Z of FIG. 7 (a), it becomes as shown in FIG. 7 (e). Since the comparison result is all true, this trader is extracted and introduced to the user B. In this way, by setting the trader attribute, general users are not extracted, so only the trader is targeted for introduction.

Although one embodiment of the present invention has been described above, it goes without saying that the present invention is not limited to the case described here.

[0031]

As described above, according to the present invention, as item attribute filtering, filtering corresponding to each user, common condition filtering,
And its mixed filtering can be realized within a fixed number of filters. Also, by using a common filter, filtering itself can be realized simply,
Moreover, flexible filter conditions can be set.

[Brief description of drawings]

FIG. 1 is a system configuration diagram of an embodiment of the present invention.

FIG. 2 is a flowchart of an embodiment of a filtering process according to the present invention.

FIG. 3 is a diagram illustrating a processing outline of the present invention.

FIG. 4 is a diagram showing a specific example of attribute divisions of attribute data.

FIG. 5 is a diagram showing a specific processing example of the present invention.

FIG. 6 is a diagram showing another specific processing example of the present invention.

FIG. 7 is a diagram showing still another specific processing example of the present invention.

[Explanation of symbols]

100 user terminal 200 Item attribute filtering device 210 Filter value management unit 220 Item Information Management Department 230 Filtering execution unit

─────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP 2000-322445 (JP, A) JP 10-207957 (JP, A) JP 8-95999 (JP, A) JP 6-215044 (JP, A) JP 2000-322262 (JP, A) JP 11-85769 (JP, A) JP 9-326792 (JP, A) JP 9-319706 (JP, A) JP Flat 11-306193 (JP, A) (58) Fields surveyed (Int.Cl. ⁷ , DB name) G06F 17/30 JISC file (JOIS)

Claims (5)

(57) [Claims] 1. Each item is divided into a plurality of attribute categories
Item management that manages items with detailed attribute data
The item group in the database is transferred to the data processing device.
More automatic filtering based on the attribute data
It is an item attribute filtering processing method, and it corresponds to each of a plurality of predetermined attribute categories
Each of the multiple attribute filter conditions
Bit correspondence with attribute data of a specific item to be processed
AND the result with the original relevant attribute filter condition.
In other words, multiple specific item attribute value set filter conditions
To create a filtering target in the item management database
For each item, the attribute data of
Constant item attribute value set Compare with each filter condition
And output items that meet all the filter conditions.
That the item attribute filtering direction, characterized in that
Law. 2. The item attribute filter according to claim 1.
In the processing method, the plurality of attribute filter conditions include each attribute filter condition.
For each attribute filter condition and the attribute data of a specific item.
Merge decision indicating whether to perform logical AND operation with data
A flag is added, and the merge determination of the plurality of attribute filter conditions is performed.
The flag performs a logical AND operation with the attribute data of a specific item.
For attribute filter conditions that indicate what to do,
Performs a logical product operation with the attribute data of a specific item
Replace the result with the original applicable attribute filter condition,
Attribute filter conditions other than
Item attribute value processing method characterized by generating an item attribute value set filter condition
Law. 3. Each item is divided into a plurality of attribute categories
Items that manage a group of items that have different attribute data
A management database, a filter management database that manages a plurality of attribute value set filter conditions corresponding to each of a plurality of attribute classifications , one of each of the plurality of attribute filter conditions, and a filter list
Bit correspondence with attribute data of a specific item to be processed
AND the result with the original relevant attribute filter condition.
In other words, multiple specific item attribute value set filter conditions
Generate and filter in the item management database
For the item to be ringed, its attribute data
One of multiple specific item attribute value set filter conditions
Compared to one, it matches all filter conditions.
And an item attribute filter link that outputs a filter link execution unit.
Processing equipment. 4. The item attribute filtering processing device according to claim 3, wherein a plurality of attribute filter sections of the filter management database are provided.
For each attribute filter condition,
Performs the logical product operation of the item and the attribute data of a specific item
A merge determination flag indicating whether or not it should be added is added, and the filtering execution means is
In the data conditions, the merge determination flag is
Indicates that the logical product operation is performed with the attribute data of the system.
Specific items for each attribute filter condition
Performs a logical product operation with the attribute data of
Replace with the attribute filter condition, and other attribute filter conditions
Use items as they are, and set multiple specific item attribute values
Item attribute filtering processing device characterized by generating a filter condition
Place 5. A recording medium on which a program for processing the item attribute filtering method according to claim 1 or 2 by a computer is recorded.