JPH11175562A - Information retrieving device and method therefor and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To integrate retrieved results obtained from plural information sources having different data items, and to present them in a batch. SOLUTION: Plural information sources to be retrieved are held in a retrieval object holding part 101, and corresponding slot data indicating a corresponding slot among those plural information sources are held in a corresponding slot data holding part 102. An information retrieving part 104 converts a retrieval condition held in a retrieval condition holding part 103 into a retrieval condition corresponding to each of plural information sources by using the corresponding slot data held in the corresponding slot data holding part 102, and retrieves the plural information sources by using the obtained retrieval condition, and holds them in a retrieved result holding part 105. An information integrating part 106 integrates the plural retrieved results thus obtained by using the substance of each information as a unit.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information processing apparatus and method for performing information retrieval processing by associating a difference in expression format between a plurality of information sources, and a storage medium storing a control program therefor.

[0002]

2. Description of the Related Art In a conventional information retrieval apparatus for retrieving information divided into items, only predetermined slots (slots represent data representing each item) can be retrieved. That is, in order to perform an information search using another slot, the information must be searched from another information search device holding the slot.

FIG. 5 is a diagram showing an example of a searchable slot in an information processing apparatus. FIG. 6 is a diagram illustrating an example of a search condition for the information processing apparatus. In FIG. 6, a column 601 shows slots for specifying conditions,
Column 602 shows the value of the condition. FIG. 7 is a diagram illustrating an example of a search result of the information processing apparatus. In FIG. 7, each row 701, 702, 703 corresponds to each slot. Which slot each column corresponds to is shown in row 711. Each row after the row 721 corresponds to the entity of one data, and the slot value of the data is shown in the corresponding column. On the other hand, FIG. 8, FIG.
FIG. 7 is a diagram illustrating an example of a searchable slot, an example of a search condition, and an example of a search result of another information processing apparatus.

As described above, a general information retrieval device can retrieve only the slots handled by the device,
Other slots must be searched using another information search device. For example, an information search device having a slot as shown in FIG. 5 can perform a search under a search condition of "Kuramoto" as shown in FIG. 6, but an information search device having a slot as shown in FIG. "
Search using the slot cannot be performed. Some devices can search from multiple sources over a network, but even in this case,
The search conditions were specified separately, and only the respective search results were obtained. For example, for an information search device having a slot as shown in FIG. 5, a search is performed using a slot of "product name", and for an information search device having a slot as shown in FIG. It was necessary to set search conditions separately, such as performing a search in the slot.

[0005] Therefore, when the above-described information retrieval apparatus is used, the user must specify the retrieval condition a plurality of times, which causes a problem that the operation becomes cumbersome.

As a technique for solving this problem, in order to process information from a plurality of information sources, corresponding slot data in which corresponding slots of frames between information sources are associated with each other is manually created, and the corresponding slot data is created. It has been proposed to make it possible to search for information expressed in different frames by utilizing the information. For example, regarding information sources having slots as shown in FIG. 5 and FIG. 8, "product name" and "brand"
The information search is performed by referring to the corresponding slot data as shown in FIG. 4 in which "price" and "price" are associated with each other. In this way, by simply instructing a search with "product name xxx", an information source having the slot of FIG. 5 performs a search by "product name xxx" and has a slot of FIG. In the information source, the search can be performed by “brand XXX”.

[0007]

However, even if data can be searched from different information sources using the corresponding slot data, basically, only the search process is individually performed for each information source. Therefore, the search results are obtained separately and there is a problem that it becomes very difficult to see.

Further, in the example of the information retrieval apparatus having the slots shown in FIGS. 5 and 8, if the user wants to check the storage and the capacity of the product name "XXX", 5 is searched using an information search device capable of searching for slots, and “Kuramoto” is obtained. Next, a search is performed using an information search device capable of searching for slots in FIG. 8 (with search conditions of “brand, xxx”). Need to get "capacity". Then, the user must look at both search results separately.

The present invention has been made in view of the above problems, and has an information retrieval apparatus and method capable of integrating search results obtained from a plurality of information sources having different data items and presenting them collectively. The purpose is to provide.

[0010] Another object of the present invention is to supplement a data item that is insufficient for one information source from another information source from a plurality of information sources having different data items, and to collect a plurality of obtained retrieval results. To be presented.

[0011]

According to one aspect of the present invention for achieving the above object, there is provided an information search apparatus for performing information search on a plurality of information sources to be searched,
The information search device has, for example, the following configuration. That is,
Holding means for holding correspondence information indicating data items corresponding to the plurality of information sources; and a given search condition, each of the plurality of information sources using the correspondence information held by the holding means A conversion unit that converts the search condition into search conditions according to the search condition, and searches the plurality of information sources based on the search conditions obtained by the conversion unit, and integrates the obtained plurality of search results with each entity of information as a unit Integration means.

According to another aspect of the present invention for achieving the above object, there is provided an information search method for performing information search on a plurality of information sources to be searched. For example, it comprises the following steps. That is, a conversion step of converting a given search condition into a search condition corresponding to each of the plurality of information sources using correspondence information indicating data items corresponding to the plurality of information sources,
An integration step of searching the plurality of information sources based on the search conditions obtained in the conversion step, and integrating the obtained plurality of search results for each entity of information.

[0013]

Embodiments of the present invention will be described below in detail with reference to the drawings.

<First Embodiment> FIG. 1 is a block diagram showing a basic functional configuration of an apparatus according to a first embodiment. In FIG. 1, reference numeral 101 denotes a search target holding unit, which holds information for specifying an information source to be searched. In this example,
The description will be made assuming that there are two information sources, a first information source as shown in FIG. 15 and a second information source as shown in FIG. Reference numeral 102 denotes a corresponding slot data holding unit.
1 is a corresponding slot data holding unit that holds information of a corresponding slot among the information sources to be searched held in 1. In this example, a description will be given assuming that corresponding slot data as shown in FIG. 4 indicating the correspondence between the slots of the first information source and the second information source is held. A search condition storage unit 103 stores search conditions. Reference numeral 104 denotes an information search unit, which outputs information from the information source specified by the information held in the search target holding unit 101 to the corresponding slot data holding unit 102.
The search condition held in the search condition holding unit 103 is converted according to each information source using the corresponding slot data held in the search condition.

Reference numeral 105 denotes a search result holding unit which holds the result searched by the information search unit 104. An information integration unit 106 integrates search results of the information held in the search result holding unit 105 using the corresponding slot data held in the corresponding slot data holding unit 102. That is, information representing the same entity is specified from search results obtained from a plurality of information sources based on the corresponding slot data, and the value of a slot missing in one information source is complemented by search results from another information source. To create integrated information. Reference numeral 107 denotes an integration result holding unit, which holds the integration information created by the information integration unit 106.

FIG. 2 is a diagram showing a specific configuration of the information processing apparatus according to the first embodiment. In the figure, 201 is C
It is a PU and operates according to a program that implements a procedure described below. Reference numeral 202 denotes a RAM, which is a search condition holding unit 1
03, the search result holding unit 105, and the integrated result holding unit 107, and provide a storage area necessary for the operation of the program. Reference numeral 203 denotes a ROM, which stores a program that implements a procedure described below. Reference numeral 204 denotes a disk device, which implements the search target holding unit 101 and the corresponding slot data holding unit 102. Reference numeral 205 denotes a network interface for acquiring information from an information source existing in another device via a network. 206 is a bus.

FIG. 3 is a flowchart for explaining the procedure of the search process according to the first embodiment. Hereinafter, the procedure of the process of the information processing apparatus according to the first embodiment will be described with reference to the flowchart of FIG. The control program for realizing the control procedure shown in FIG.
And executed by the CPU 201.

In step S301, the search target holding unit 1
Regarding the information source to be searched held in 01, it is checked whether or not an unprocessed information source remains. If the information source remains, the process proceeds to step S302. On the other hand, if it is determined in step S301 that there is no unprocessed information source, this processing ends.

In step S302, the search target holding unit 1
One unprocessed information source is selected from among the information sources to be searched held in 01. For example, first, the first information source is selected. Then, control goes to a step S303. Step S3
In step 03, the search condition held in the search condition holding unit 103 is converted according to the information source selected in step S302. Specifically, the corresponding slot data holding unit 10
The slot of the search condition is converted into the slot of the selected information source by using the information of the corresponding slot held in 2. Then, control goes to a step S304.

FIG. 4 is a diagram showing an example of corresponding slot data held in the corresponding slot data holding unit 102 according to the first embodiment. Here, columns 401 and 402
Respectively correspond to a first information source and a second information source. Row 4
11 and 412 hold corresponding slots between the information sources. Further, as described above, the search target holding unit 101 has the first information source (FIG. 15) having a slot as shown in FIG. 5 and the second information source (FIG. 8) having a slot as shown in FIG. FIG. 16) is held.

Now, the corresponding slot data held in the corresponding slot data holding unit 102 is as shown in FIG. 4, and the search condition holding unit 103 stores the search conditions shown in FIG. Is held. In this state, the search condition shown in FIG.
When the conversion for performing the search in the first information source capable of searching for the slot indicated by is performed, the search condition (“product name OO”) shown in FIG. 11 is obtained.

In step S304, step S303
The search is performed from the selected information source (here, the first information source) using the search conditions converted in step (1), and the search result is stored in the search result storage unit 105. The search result in this case is shown in FIG.
As shown in FIG. Then, the process proceeds to step S305.

After step S305, an integration process is performed on the search results. First, in step S305, it is checked whether or not unprocessed search results remain in the search result holding unit 105.
Move on to In step S306, the search result holding unit 105
One unprocessed search result is selected from the search results stored in. Then, control goes to a step S307. Step S
In 307, information indicating the same entity as the result selected in step S306 is identified from the results held in the integration result holding unit 107. For example, by referring to the corresponding slot held in the corresponding slot data holding unit 102, a corresponding slot between the result selected in step S306 and the result held in the integration result holding unit 107 is obtained, Those corresponding values of the corresponding slots are identified as information indicating the same entity. If there is an identified result, the process proceeds to step S308. If there is no identified result, the process proceeds to step S309. In step S309, the integration result storage unit 107 stores the search result as it is.

If there is an identified result, that is, if there is a search result representing the same entity, step S308
, The search results related to the same entity are integrated. In this integration process, a slot that does not exist in the integration result identified as the same entity but exists in the search result is extracted, and when such a slot is extracted, it is added to the integration result. . Then, the process returns to step S305. In step S309, the searched result is
The result is stored in the integration result storage unit 107 according to the corresponding slot. Then, the process returns to step S305.

When the processing of steps S306 to S309 has been completed for all the search results for the information source, the process returns from step S305 to step S301, and the processing is performed for the next information source. For example, when the processing is completed for all the search results of the first information source, the process returns to step S301, the search processing of "brand OO" is performed for the second information source, and the integration processing shown in steps S306 to S309 is performed.

FIG. 12 is a diagram showing a result of integrating search results by two information sources according to the first embodiment. In the above example, first, the first information source is searched for “product name, xxx”, and FIG. 12A is obtained as a search result. At this time, by the integration processing of steps S306 to S309, the search result of FIG. Next, the second information source is “brand, xx
The search result shown in FIG. 12B is obtained. Here, the search result of FIG. 12A and the search result of FIG.
When the search results are compared, the slots of [Product name and brand] registered in the corresponding slot data of FIG.
Since both “price and price” are 4000 yen, both search results are identified as indicating the same entity (S307). Then, in step S308, FIG.
(A) does not exist in the search result (that is, the integration result),
The slot “capacity” existing in the search result of FIG. 12B is extracted and added to the integrated result. As a result, an integrated result as shown in FIG. in this way,
In the past, the "capacity" and "Kuramoto" slots could only be obtained separately for each search result by simply searching for each information source from the search condition "brand xxx". According to FIG. 12, it is possible to obtain a result obtained by integrating these slots as shown in FIG.

In the above embodiment, a similarity as shown in a third embodiment described later can be used as a method of determining a slot for identifying whether or not the entity is the same entity.

<Second Embodiment> In the first embodiment described above, a case will be described in which the operation is performed on the premise that the corresponding slot data holding unit 102 holds information of a separately created corresponding slot. However, the present invention is not limited to this. In the second embodiment, a configuration in which corresponding slot information between information sources held in the search target holding unit 101 is automatically created will be described.

FIG. 13 is a block diagram showing a basic functional configuration of the device according to the second embodiment. In FIG. 13, the same components as those in the first embodiment (FIG. 1) are denoted by the same reference numerals, and description thereof will be omitted. Reference numeral 110 denotes a corresponding slot data creating unit that creates corresponding slot data indicating a corresponding slot among the information sources held in the search target holding unit 101 and causes the corresponding slot data holding unit 102 to hold the corresponding slot data.

The procedure for creating the corresponding slot data by the corresponding slot data creating section 110 will be described below. FIG. 14 is a flowchart for explaining the procedure for creating the corresponding slot data.

First, in step S31, one frame held in the frame definition holding unit 101 is denoted by f, the other frame is denoted by f ', and a set of slots of the frame f is represented by Sf (f). Hypothesis Ps = {(s, s ') | s {Sf (f), s'} S where slots of both frames are arbitrarily combined
f (f ′)｝ is generated, and a hypothesis set H in which hypotheses are collected is created.

For example, for the frames shown in FIGS. 5 and 8, the contents of H are as follows. In this case, 1-3
There are a pair of slot pairs.

First, there are the following nine combinations as one set of slot pairs. That is, {(product name, brand)}, {(product name, price)}, {(product name, capacity)}, {(price, brand)}, {(price, price)},
｛(Price, capacity)｝, ｛(Kuramoto, brand)｝, ｛(Kuramoto,
Price)｝, ｛(brewery, capacity)｝.

As the two slot pairs, there are the following 18 combinations. That is, ｛(product name,
(Brand), (price, price)｝, ｛(brand name, brand), (price, capacity)｝, ｛(brand name, price), (price, brand)｝, ｛(brand name, price), (price) ,capacity)},
｛(Product name, capacity), (price, brand)｝, ｛(product name,
(Capacity), (Price, Price)｝, ｛(Product Name, Brand), (Kuramoto, Price)｝, ｛(Product Name, Brand), (Kuramoto, Capacity)｝, ｛(Product Name, Price), (Kuramoto , Brand)｝,
｛(Product name, price), (brewery, capacity)｝, ｛(product name,
(Capacity), (Kuramoto, Brand)｝, ｛(Product Name, Capacity), (Kuramoto, Price)｝, ｛(Price, Brand), (Kuramoto, Price)｝,
｛(Price, Brand), (Kuramoto, Capacity)｝, ｛(Price, Price), (Kuramoto, Brand)｝, ｛(Price, Price), (Kuramoto,
(Capacity)｝, 容量 (Price, Capacity), (Kuramoto, Brand)｝,
｛(Price, capacity), (brewery, price)｝.

As the three slot pairs, there are the following six combinations. That is, {(product name, brand), (price, price), (brewery, volume)}, {(product name, brand), (price, volume), (brewery, price)},
｛(Product name, price), (price, brand), (brewery, capacity)｝, ｛(product name, price), (price, capacity), (brewery, brand)｝, ｛(product name, capacity), (Price, brand),
(Kuramoto, price)｝, {(product name, capacity), (price, price), (Kuramoto, brand)}.

As described above, in step S31, 33
A type of hypothesis set H is obtained. Then, step S32
Move on to

In step S32, when a set of slot pairs for one frame is selected from the set of slots constituting the hypothesis from the set of hypotheses H obtained above, the combination of the slot values is stored in the information source. Information (below,
Only the hypotheses that are all different from each other are called.

Hereinafter, step S32 will be specifically described with reference to FIGS. In particular, for one set of slots, {(product name, brand) obtained in step S31}, and for two sets of slots, {(product name, brand), (price, price) obtained in step S31. ｝ And ｛(price, brand), (brewery, price)｝ For three sets of slots, {(product name, brand), (price, price), (brewery, capacity)} obtained in step S31 It will be described using FIG.

First, for one set of slot pairs {(product name, brand)}, it is determined whether or not the values of 1301 to 1304 in FIG. 15 are all different depending on the slot value of the product name (one frame). Looking at the product names in FIG. 15, they are OO, △△△, ☆, ▽▽▽, and ▽▽▽, which are all different. Also, regarding the slot value of the brand in the other frame, if the comparison is made in the same manner as in FIG. 16, all are different, so that {(product name, brand)} remains as a hypothesis. Hereinafter, if the same comparison is made for the other “one set of slots”, the hypothesis of {(product name, brand)} remains.

In addition, two sets of slots {(product name, brand), (price, price)} may be compared in the same manner as one set of slots. In the case of two sets, the product name and price, Alternatively, the value obtained by combining both the brand and the price needs to be different in all of 1301 to 1304 in FIG. In this case, in the combination of the merchandise name and the price slot value, the prices of 1303 and 1304 are the same value at 6000 yen, but the merchandise names are different between ☆ and ▽▽▽. Similarly, when examining combinations of brand and price slot values, all the combination values are different in 1401 to 1403 in FIG. Therefore, ｛(product name, brand),
The hypothesis of (price, price)｝ remains.

On the other hand, when examining {(product name, capacity), (price, price)}, the combination of the capacity and the price slot value is the same (capacity: 720 ml, price: 400).
Since two (0 yen) are shown in FIG. 16 (1401, 1402), the information is not uniquely determined, and {(price, brand), (brewery, price)} does not remain as a hypothesis.

Similarly, when the other two sets of slots are examined, {(product name, brand), (price, price)}, {(product name, brand), (price, capacity)},
｛(Product name, price), (price, brand)｝, ｛(product name,
(Capacity), (Price, Brand)｝, ｛(Product Name, Brand), (Kuramoto, Price)｝, ｛(Product Name, Brand), (Kuramoto, Capacity)｝, ｛(Product Name, Price), (Kuramoto , Brand)｝,
The hypothesis of {(product name, capacity), (brewery, brand)} remains.

For three sets of slots {(product name, brand), (price, price), (brewery, capacity)},
Combination of product name, price and brewery slot value is shown in FIG.
1301 to 1304 are different, and the combination of the brand, price, and slot value of the capacity is 1401 in FIG.
~ 1403 are all different. Therefore, ｛(product name,
(Brand), (price, price), (brewery, capacity)｝ hypothesis will remain.

After all, in the case of this example, with respect to three sets of slots, all of the above six combinations remain as hypotheses.

As described above, in step S32,
Processing is performed to leave only hypotheses for which information is uniquely determined.
Then, in step S33, the process proceeds to step S33. For all the hypotheses in the set H, an instance having the same slot value is obtained in every slot pair in the hypothesis, and the instance is set as a hypothesis instance pair set. For example,
In {(product name, brand), (price, price), (brewery, capacity)}, the slot pair of (product name, brand)
OO and ☆☆☆ exist in both sources and are the matching instances. Also, in the (price, capacity) slot pair, they match at 4000 yen and 6000 yen.
However, there is no matching value in the (storage, capacity) slot pair.

A set having a slot pair in which no matching instance exists, such as (stock, capacity), is removed from the hypothesis set H. As a result, the remaining hypothesis is ｛(product name,
(Brand)｝ {(product name, brand), (price, price)}. Then, the process proceeds to step S34.

In step S34, the hypothesis having the same number of slot pairs in the hypothesis and the largest number of instance pairs found is taken. If there is more than one with the maximum value,
For example, take the hypothesis generated earlier. These are arranged in K in ascending order of the number of elements (the number of pairs of slot pairs). In the case of the present example, in step S33, only one hypothesis remains in each of one set of slot pairs and two sets of slot pairs, so the hypotheses left in step S33 are arranged in K as they are.

Next, in step S35, it is checked whether or not a hypothesis remains in K, and if it remains, the process proceeds to step S36. Here, if no hypothesis remains in K, all the processing ends. In this case, the corresponding slot data cannot be obtained.

In step S36, the j-th hypothesis h of K
When the number of instance pairs obtained for j is nj,
The maximum hj satisfying nj-1 / nj <1/2 is defined as the corresponding slot data. In this way, false slot pairs are reduced. If there is no such hypothesis, j = 1.

Then, the process proceeds to a step S37. In step S37, the slot pair of hj is held in the corresponding slot pair data holding unit 104 as the corresponding slot pair data. Then, all processing ends.

In the above example, {(product name, brand), (price, price)} remains as a final hypothesis, and corresponding slot data as shown in FIG. 4 is obtained.

As described above, there is an effect that the processing of the information of the information source described in the different frame format can be automatically executed without manually giving the corresponding slot data.

Third Embodiment In the first embodiment, a case has been described in which the search conditions held in the search condition holding unit 103 are converted and only information that can be directly searched is integrated. However, the present invention is not limited to this.
Using the searched result, information about the same entity as the result can be searched from another information source and integrated.

FIG. 17 is a block diagram showing a basic functional configuration of the device according to the third embodiment. 17, the same components as those in the first embodiment (FIG. 1) are denoted by the same reference numerals, and description thereof will be omitted. Reference numeral 120 denotes a re-search unit. The re-search unit 120 refers to the result held in the integration result holding unit 107 and searches the information source that cannot be directly searched by the search condition held in the search condition holding unit 103. Create and search again.

With the search conditions held in the search condition holding unit 103, not all information sources held in the search object holding unit 101 can be searched. If there is no slot corresponding to the information source, it cannot be searched. For example, it is assumed that the search condition shown in FIG. At this time, the information source that can search for the slot shown in FIG. 8 cannot be searched because there is no slot corresponding to “Kuramoto”. Therefore, after one kind of integration result is obtained by the search condition of FIG. 6, a new search condition is created from the information of the integration result, and the search is performed again from the information source that could not be searched to integrate the information.

In the above example, the search condition “Kuramoto, ×
Xx brewing ”is searched for the information source of FIG.
The result shown in FIG. Therefore, "Product name" is changed to "XX
If a re-search is performed under the condition of "○" and the condition of "product name" is "△△△", "product name (brand)" is changed to "○○○" from the second information source in FIG. A search result as shown in FIG. 10 is obtained for a certain entity. And
From the search results in FIG. 10, for brands with
Another information "capacity" is obtained. From this result, if the information integration processing described in steps S305 to S309 is performed, the integration result shown in FIG. 18 is obtained.

In the third embodiment, a search condition for performing a search in another information source is automatically generated from the obtained integration result. Since the search condition adds the search result to the already obtained integration result, it is necessary to set a search condition capable of identifying the same entity as the entity of the information represented by the integration result.

As a method for generating such a search condition, 2
It is explained as follows.

First, as a first method, of the slots of the obtained integration result, the slots in which the corresponding slots are set in the corresponding slot data are extracted, and AND of these slots is set as a search condition. can give. For example, in the search result of FIG. 7, there are three slots of “product name”, “price”, and “brewery”, and among these slots, the corresponding slot data of FIG.
And "price" exist. Therefore, in the re-search unit 120,
Using the AND of "Product Name" and "Price", "(Product Name:
○○○) & (Price: 4000 yen) ”and“ (Product name:
Δ △△) & (price: 4000 yen) ".

As a second method, a uniqueness is defined for each slot, and a slot having a uniqueness equal to or more than a certain value is extracted from the integration result as a search condition. Here, the uniqueness uniq (s, r) of the slot s of the information source r can be defined as follows, for example.

Now, the probability p (ik) of occurrence of each instance ik of the instance set Ir of the information source r is all 1 / n
(Ir). Here, the elements of the set A are represented by n
(A).

At this time, the X average information amount of the event that occurs in each instance of Ir is equal to the entropy H (X),

[0063]

(Equation 1)

Is as follows.

Next, for the slot s, the set V is defined as V = {va1 (s, ik) | 1 ≦ k ≦ n (Ir)}. Here, va1 (s, i) is an instance i
Is the value of the slot s. The j-th element of V is vVj,
Assuming that the number of instances ik satisfying vk = vVj in Ir is cj, the probability of occurrence of the value vVj is p (vVj) = cj / n (Ir).

Therefore, the mutual information I (X; Y) between the event Y and the event X at which the value V occurs is the probability p (vVj, ik) at which vVj and ik occur simultaneously.

[0067]

(Equation 2)

Considering that

[0069]

(Equation 3)

Is obtained.

With the above preparation, the uniqueness uniq (s, r) of the slot s in the information source r is represented by i (X; Y)
/ H (X). That is, it becomes as follows.

[0072]

(Equation 4)

As is clear from the definition, in the information source r, for example, the uniqueness becomes 1 when the values of the slots s are all different, and becomes 0 when they are all the same. Therefore, a new search condition is generated from the integration result using the slot having the uniqueness of 1. In the above example, "(Product name: XX
○) ”and“ (product name: △△△) ”are generated as new search conditions. Although the similarity is calculated for one slot s, the similarity can be similarly calculated for a plurality of sets of slots.
In the third embodiment, a set of a plurality of slots can be used as a new search condition.

[Other Embodiments] In the processing described with reference to FIG. 3, information indicating the same entity is identified at the time of integrating search results. However, the present invention is not limited to this. The information indicating the entity of may be obtained. For example, a unique number may be assigned to each instance, and a set of instance numbers indicating the same entity may be held. Such a set of instances can be obtained in advance using the information of the corresponding slot obtained in the second embodiment.

In FIG. 3, a case where all slots are integrated is described. However, the present invention is not limited to this case, and integration may be performed only for a previously specified slot.

FIG. 3 illustrates the case where the integration of information is performed for each search, but the present invention is not limited to this. After all the searches are completed, the integration process may be performed collectively.

In FIG. 3, the case where the search process and the integration process are sequentially performed has been described, but the present invention is not limited to this. Processing may be performed in parallel,
Some search processing and integration processing may be partially performed collectively. For example, the third search may be executed while integrating the first and second search results. Further, when there are information sources A, B, C, and D, the information sources A and B are searched for an integrated result E, the information sources C and D are searched for an integrated result F, and finally the integrated result is obtained. E and F may be integrated.

In the above embodiment, the corresponding slot used for identifying information indicating the same entity is:
One set or a plurality of sets may be used. Further, a candidate for a slot used for identification may be determined in advance.

In the above embodiment, when the slot values completely match, the information is identified as the same entity. However, the present invention is not limited to this, and the similarity (similarity) of the slot values is determined. May be identified as pointing to the same entity.

When such a similarity is used, if the score calculated from the similarity is a certain value or more, it is identified as the same information. As a method of calculating the score s, for example,
The following equation can be used:

[0081]

(Equation 5)

Here, υ and υ ′ represent the value of each slot. Similarity sim of slot values υ and υ '
For example, (υ, 、 ′) is obtained by normalizing the score of the optimal collation search described in “Natural Language Processing”, edited by Makoto Nagao, Iwanami Shoten, 1996, pp. 431 to 434, between 0 and 1. Use something.

Further, in the above description, the case where the optimal collation search that produces a measure similar to a character string is used as the similarity of the slot value has been described. However, the present invention is not limited to this. You may. For example, for a slot value composed of numbers, a method of using the similarity as a numerical value can be considered. In addition, the similarity between the slot value of the alphabet and the slot value of the katakana is obtained by converting an English word or the like written in the alphabet into katakana notation (for example, IPSJ Journal Vol. 35, No. 1, Sumiyoshi,
There is a method described in Aizawa "Katakana conversion of English proper nouns"), and a method of using the similarity between the converted katakana character string and the other katakana character string. Alternatively, there is a method in which the similarity is higher when the distance on the thesaurus is shorter, using a thesaurus.

Further, there is a method in which all of the plurality of similarities are obtained, and the highest value is employed.

Further, in the above embodiment, the case where each unit is configured on the same computer has been described. However, the present invention is not limited to this, and each unit is divided into computers and processing devices distributed on a network. You may comprise.

In the above embodiment, the case where the search condition holding unit 103, the search result holding unit 105, and the integration result holding unit 107 are implemented by a RAM and the corresponding slot data holding unit 102 is implemented by a disk device has been described. The present invention is not limited to this, and may be realized using an arbitrary storage medium. The same applies to other embodiments.

In the above embodiment, the case where the program is stored in the ROM has been described. However, the present invention is not limited to this, and may be realized by using an arbitrary storage medium. Further, it may be realized by a circuit that performs the same operation.

The present invention may be applied to a system composed of a plurality of devices or an apparatus composed of a single device. A recording medium storing the software program code for realizing the functions of the above-described embodiments is supplied to a system or an apparatus, and a computer (or CPU or MPU) of the system or the apparatus reads out the program code stored in the recording medium. Needless to say, it can also be achieved by executing.

In this case, the program code itself read from the recording medium implements the functions of the above-described embodiment, and the recording medium on which the program code is recorded constitutes the present invention.

Examples of a recording medium for supplying the program code include a floppy disk, hard disk, optical disk, magneto-optical disk, CD-ROM,
DR, magnetic tape, nonvolatile memory card, ROM
Etc. can be used. When the computer executes the readout program code, not only the functions of the above-described embodiments are realized, but also the OS or the like running on the computer performs actual processing based on the instruction of the program code. Do some or all,
It goes without saying that a case where the function of the above-described embodiment is realized by the processing is also included.

Further, after the program code read from the recording medium is written into a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, based on the instruction of the program code, It goes without saying that the CPU provided in the function expansion board or the function expansion unit performs a part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

[0092]

As described above, according to the present invention, search results obtained from a plurality of information sources having different data items can be integrated and presented together. For this reason, the user can conveniently view the search results of a plurality of information sources.

[0093]

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a basic functional configuration of an apparatus according to a first embodiment.

FIG. 2 is a diagram illustrating a specific configuration of the information processing apparatus according to the first embodiment.

FIG. 3 is a flowchart illustrating a procedure of a search process according to the first embodiment.

FIG. 4 is a diagram showing an example of corresponding slot data held in a corresponding slot data holding unit 102 of the first information source according to the first embodiment.

FIG. 5 is a diagram showing slots of a first information source according to the first embodiment.

FIG. 6 is a diagram showing an example of a search condition.

FIG. 7 is a diagram illustrating an example of a search result of a first information source.

FIG. 8 is a diagram illustrating slots of a second information source according to the first embodiment.

FIG. 9 is a diagram illustrating an example of a search condition.

FIG. 10 is a diagram illustrating an example of a search result of a second information source.

FIG. 11 is a diagram showing search conditions.

FIG. 12 is a diagram showing a result of integrating search results by two information sources according to the first embodiment.

FIG. 13 is a block diagram illustrating a basic functional configuration of an apparatus according to a second embodiment.

FIG. 14 is a flowchart illustrating a procedure for creating corresponding slot data.

FIG. 15 is a diagram illustrating contents of a first information source according to the second embodiment.

FIG. 16 is a diagram illustrating contents of a second information source according to the second embodiment.

FIG. 17 is a block diagram illustrating a basic functional configuration of an apparatus according to a third embodiment.

FIG. 18 is a diagram illustrating a result of integrating search results of two information sources according to the third embodiment.

[Explanation of symbols]

 101 search target holding unit 102 corresponding slot data holding unit 103 search condition holding unit 104 information search unit 105 search result holding unit 106 information integrating unit 107 integrated result holding unit

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Noriko Otani 3-30-2 Shimomaruko, Ota-ku, Tokyo Inside Canon Inc.

Claims (19)

[Claims] 1. An information search apparatus for performing information search on a plurality of information sources to be searched, comprising: holding means for holding corresponding information indicating data items corresponding to the plurality of information sources; Conversion means for converting a search condition into search conditions corresponding to each of the plurality of information sources using the correspondence information held by the holding means; and And an integrating means for integrating the plurality of obtained search results for each information entity as a unit. 2. The apparatus according to claim 1, wherein the integration unit identifies information on the same entity from the plurality of search results, and integrates information identified as the same entity. Information retrieval device. 3. The integration unit according to claim 1, wherein the search results obtained from the two information sources have matching values in the data items for which correspondence is set in the correspondence information held in the holding unit. 3. The information search apparatus according to claim 2, wherein the two search results are identified as information of the same entity. 4. A generating means for generating a new search condition from a search result obtained by said integrating means, and a re-searching means for performing a search again for said plurality of information sources in accordance with the search condition generated by said generating means. The information retrieval apparatus according to claim 1, further comprising: 5. The information search apparatus according to claim 4, wherein the generation unit generates a search condition for uniquely specifying an entity of information based on a search result obtained by the integration unit. . 6. The generation unit extracts a data item held in the correspondence information from among data items present in the search result obtained by the integration unit, and generates a new product by logical product of the extracted data items. The information search device according to claim 5, wherein a search condition is generated. 7. The generating means calculates uniqueness for each data item of the search result obtained by the integrating means, extracts data items in which each entity of information has unique data, and extracts the extracted data items. The information search device according to claim 5, wherein a new search condition is generated based on the data item. 8. The information retrieval apparatus according to claim 1, wherein said information integration means performs integration only for a specific data item specified in advance. 9. Based on the data of the plurality of information sources,
9. The apparatus according to claim 1, further comprising correspondence information generation means for generating correspondence information to be held by said holding means.
An information retrieval device according to any one of the above. 10. An information search method for performing an information search on a plurality of information sources to be searched, wherein a given search condition is determined using correspondence information indicating data items corresponding to the plurality of information sources. A conversion step of converting to a search condition corresponding to each of the plurality of information sources; and searching the plurality of information sources using the search conditions obtained in the conversion step, and obtaining a plurality of search results obtained by each information. An integration step of integrating the entities of the information as a unit. 11. The method according to claim 10, wherein the integrating step identifies information on the same entity from the plurality of search results, and integrates information identified as the same entity. Information retrieval method. 12. The integration step according to claim 1, wherein the search results obtained from the two information sources have matching values in the data items for which correspondence is set in the correspondence information held in the holding step. And identifying the two search results as information of the same entity.
Information search method described in. 13. A generating step of generating a new search condition from a search result obtained in the integrating step, and a re-searching step of performing a search again for the plurality of information sources according to the search condition generated in the generating step. 13. The information search method according to claim 10, further comprising: 14. The information search method according to claim 13, wherein the generating step generates a search condition for uniquely specifying an entity of information based on a search result obtained in the integrating step. . 15. The generating step extracts a data item held in the correspondence information from data items present in the search result obtained in the integrating step, and creates a new logical product of the extracted data items. The information search method according to claim 14, wherein a search condition is generated. 16. The generating step calculates a uniqueness for each data item of the search result obtained in the integrating step, extracts a data item in which each entity of information has only one data, and extracts the extracted data item. The information search method according to claim 14, wherein a new search condition is generated based on the data item. 17. The information search method according to claim 10, wherein in the information integration step, the integration is performed only for a specific data item specified in advance. 18. The apparatus according to claim 10, further comprising a correspondence information generating step of generating correspondence information to be held in said holding means based on data of said plurality of information sources. Information search method described. 19. A storage medium for storing a control program for causing a computer to perform information search on a plurality of information sources to be searched, wherein the control program stores a given search condition for a plurality of information sources. A code for a conversion step of converting the information into search conditions corresponding to each of the plurality of information sources using the corresponding information indicating the corresponding data items; and searching the plurality of information sources using the search conditions obtained in the conversion step And a code for an integration step of integrating a plurality of obtained search results in units of each entity of information.