Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1999-12-20
2002-10-22
Von Buhr, Maria N. (Department: 2125)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C707S793000
Reexamination Certificate
active
06470337
ABSTRACT:
BACKGROUNDS OF THE INVENTION
1. Field of the Invention
The present invention relates to an information retrieval system, and more particularly to an information retrieval system that can narrow down a retrieval result by use of hierarchical index of a subject to be retrieved and its method, and a storing medium with an information retrieval program stored therein.
2. Description of the Related Art
When retrieving desired information through access to a database, a user inputs some retrieval condition so to get the retrieval result. When there exist a lot of retrieval results, he or she further enters an additional retrieval condition, depending on necessity, so to narrow the search.
In the information retrieval by the conventional information retrieval system, since any retrieval condition is entered by a user, he or she would repeat trial and error while adding a retrieval condition or deleting it, until getting a properly-narrowed retrieval result.
Then, some techniques for supporting decision of a retrieval condition in the information retrieval have been proposed so far. As this kind of technique, there are a technique of setting a key word index to each information that is disclosed in the Japanese Patent Publication Laid-Open (Kokai) No. Heisei 7-65020, and an information retrieval system of presenting retrieval conditions for narrowing a search by a hierarchical index, to a user, that is disclosed in the Japanese Patent Publication Laid-Open (Kokai) No. Heisei 4-114277.
These conventional techniques for supporting decision of a retrieval condition, however, are only to present an index means established based on the information stored in a database that is a subject to be retrieved, and they have no means for presenting retrieval conditions one after another from the most efficient condition for narrowing a retrieval result according to the retrieval probability of information and a hierarchical index.
As mentioned above, the conventional information retrieval system has no means for supporting decision of a retrieval condition to be added for narrowing a retrieval result, or, if it has, the conventional system cannot present retrieval conditions one after another from the most efficient condition for narrowing a retrieval result based on the retrieval probability of the information.
Therefore, when there is no means for supporting decision of a retrieval condition, a user has to select a proper retrieval condition while repeating trial and error, and even when there is the means for supporting decision of a retrieval condition, the conventional system cannot decide a proper retrieval condition efficiently.
SUMMARY OF THE INVENTION
In order to solve the above mentioned problem, the present invention aims to provide an information retrieval system capable of deciding a proper retrieval condition efficiently and narrowing down a search into a proper result by presenting nodes of hierarchical index in the most effective order for narrowing down a retrieval result, and provide its method and a storing medium with such an information retrieval program stored therein.
According to the first aspect of the invention, an information retrieval system for retrieving desired data from a retrieved subject database with data to be retrieved stored therein, comprises
retrieval condition expression creating means for creating a retrieval condition expression by use of an entered retrieval condition,
retrieval executing means for executing retrieval processing through access to the retrieved subject database based on a retrieval condition expression created by said retrieval condition expression creating means,
index storing means for storing each hierarchical index as for retrieved subject data stored in the retrieved subject database,
retrieved subject retrieval probability storing means for storing record of retrieval times of each retrieved subject data stored in the retrieved subject database, and
narrowing means of obtaining the index corresponding to the retrieved subject data obtained by said retrieval executing means as a retrieval result, from said index storing means, computing expected value of acquired information based on the retrieval times of the retrieval result data stored in said retrieved subject retrieval probability storing means, and presenting to a user the index information corresponding to the retrieval results, in decreasing order of the expected value of the computed acquired information as for the retrieved subject data that is the retrieval result, for narrowing down a retrieval result obtained by said retrieval executing means, in reply to a user's narrowing request.
In the preferred construction, the information retrieval system further comprises retrieval condition receiving means for receiving a user's input of a retrieval condition, and retrieval result display means for displaying a retrieval result obtained from the retrieval processing by said retrieval executing means.
In another preferred construction, said narrowing means obtains, from the hierarchical index stored in said index storing means, a question node that is a node positioned at the lowest hierarchy, of the common nodes in all the paths, and an answer node that is a node positioned at a lower hierarchy right beneath the question node, assuming a path starting from a leaf indexed to the retrieved subject data obtained by said retrieval executing means as a retrieval result and tracing the nodes of hierarchical index up to the upper hierarchy, computes the expected value of the acquired information as for the respective obtained question nodes, and presents to a user the question nodes in decreasing order of the expected value of the computed acquired information and the answer nodes corresponding to the respective question nodes, and further
said narrowing means computes information amount M(C) of collection C of the retrieved subject data by the following expression, assuming that the collection of the retrieved subject data that is a retrieval result by said retrieval executing means is defined as C, by use of the number k of the retrieved subject data that is the retrieval result and the existence ratio Pj of the retrieved subject data j of the collection C of the retrieved subject data,
M
(
C
)
=
-
∑
j
=
1
k
P
j
log
2
P
j
computes the expected information amount B (C, a) when the collection C is divided into partial collections C
1
, . . . , Cn based on the answer nodes a
1
, . . . , an as for the given question node a, by the following expression,
B
(
C
,
a
)
=
∑
i
=
1
n
|
C
i
|
|
C
|
M
(
C
i
)
and computes the expected value gain (C, a) of the acquired information of the respective question nodes, by the following expression.
gain(
C,a
)=
M
(
C
)−
B
(
C,a
)
In another preferred construction, said narrowing means makes the existence ratio Pj of the retrieved subject data j in the collection C, as the retrieval probability pj of the retrieved subject data rj that is computed by the following expression, by use of the number k of the retrieved subject data that is a retrieval result by said retrieval executing means, the retrieved subject data r
1
, r
2
, . . . , rm that is the retrieval result, the retrieval times hj of the given retrieved subject data rj (1≦j≦m), and the number of times vj the retrieval result appears at a lower hierarchy below a different answer node.
p
j
=
h
j
v
j
∑
i
=
1
k
h
j
In another preferred construction, said narrowing means obtains, from the hierarchical index stored in said index storing means, a question node that is a node positioned at the lowest hierarchy, of the common nodes in all the paths, and an answer node that is a node positioned at a lower hierarchy right beneath the question node, assuming a path starting from a leaf indexed to the retrieved subject data obtained by said retrieval executing means as a retrieval result and tracing the nodes of hierarchical index up to the upper hierarchy, computes the expected value of the acquired information as fo
NEC Corporation
Von Buhr Maria N.
LandOfFree
Information retrieval system using a hierarchical index for... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Information retrieval system using a hierarchical index for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Information retrieval system using a hierarchical index for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2993051