JPS6151238A - Accessory device to retrieved information - Google Patents

Abstract

PURPOSE:To attain automatic provision of a mark minimizing a retrieval time when the mark is given to a coded document as retrieval hint information by providing an optimum value calculation/optimum mark extracting means or the like. CONSTITUTION:The optimum value calculation/optimum mark extracting means 5 calculates which mark is to be given for optimization based on a statistic information storage means 2, an optimum evaluation type storage means 3 and a ratio command information storage means 4 and extracts a proper mark. In this case, when plural extracted marks exist, an optional mark, e.g, the mark extracted at first is selected automatically and given to a mark providing means 6. That is, the means 6 decides automatically the information to be given to each document, the mark is provided while both the document and given mark are displayed on a display. Thus, the key word is given automatically, the trouble is relieved and the autmatic provision of the mark is attained while minimizing the retrieval time.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an apparatus for adding search information to electronic documents.

(Prior Art) Conventionally, in a search system such as documents and patents, or in the management of general documents created and coded using a word processor or the like.

With the introduction of optical disk files, document storage of image data is increasing, and identification numbers and

The mainstream is to assign keywords related to the content and use this information to perform a search. In addition to this, 1
Document management device that adds color and searches
-121767) etc. are known.

(Problems with the Prior Art) However, when simply using an identification number as a search means, it is assigned mechanically, making it extremely difficult for the user to memorize the number. Furthermore, when using keywords, the specification at the time of assignment is complicated, and the user's wishes are not reflected. In other words, it is difficult to reduce both the effort required to provide search information and the effort required to perform a search. In order to solve these problems,
There are some methods, such as the above-mentioned published patent publication, that attempt to store and search based on color, but there is no mechanism to minimize search time, so the scope of application is limited to managing small-scale documents. There are problems. for example,
Red, yellow, green, '* 4 for stored documents.
Let us consider a case where a search is performed by assigning one of several types of colors. At this time, one color can be assigned freely, so if red is specified for most documents, the search based on this color assignment will not be effective, and the efficiency of the search will not be improved. This is either a fool's errand, or rather it takes time to add color information. In addition, since this search information is -dimensional, there are inherent problems such as the inability to take action when the user's memory of the search information fades. In order to manage this, it is essential to make input easy and to reduce the search effort, especially the user's search time.In addition, in order to deal with user forgetting, it is essential to It is important to construct medical examination information.

(Purpose of the invention) The first object of the present invention is to solve the above problems.

To provide a device that automatically adds a mark so as to minimize the search time when adding a mark to an electronic document to use it as search key information. Here, instead of the mark, you can replace it with general search element information that is not directly related to the content of the document, such as coloring the document or background.
In extreme cases: ``Words'' that have no direct semantic relationship to the content of the dormitory may be used. The marks written below have the same meaning.

A second object of the present invention is to provide a dialog box marking method that satisfies the first object and takes into account the user's preferences.

The third object of the present invention is to provide information on the attributes of marks (shape, color, size, etc. associated with the mark) in order to prevent users from forgetting electronic documents.
, we construct multi-dimensional trial information from multiple viewpoints, and automatically add marks that combine mark attributes to minimize search time, or multiple search elements. The objective is to provide a device that

A fourth object of the present invention is to provide an interactive mark attribute or multiple clue information adding device that takes into account the user's preferences while satisfying the third object. (Structure of the Invention) According to the present invention, in a search information adding device for adding marks to digitized documents in a document management system as search key information, there is provided means for storing a plurality of marks in advance.

means for storing statistical information of assigned marks, means for pre-storing a search optimization evaluation formula, means for pre-storing instruction information of mark assignment ratio from a user, and the statistical information and assignment ratio instruction information. , a means for calculating an optimum value using an optimization evaluation formula and extracting an appropriate mark from the mark recording means, and adding the extracted mark to the document.

It is characterized by having means for updating the corresponding statistical information of the means for storing the statistical information, and automatically determining and attaching the mark so as to minimize the search time when registering a document. A device for providing examination information is obtained.

Further, according to the present invention, in a search information adding apparatus for adding marks to a computerized document in a document management system to obtain search information, means for pre-storing a plurality of marks, and the added marks. means for storing statistical information of , means for pre-storing a search optimization evaluation formula, means for pre-storing instruction information of a mark assignment ratio from a user, and the statistical information and assignment ratio instruction information.

Means for calculating an optimum value using an optimization evaluation formula and extracting an appropriate mark from the mark storage means; Mark the candidate statistics for the Fμ number that gives the optimal value 'v#
The above-mentioned markings have a means of updating the corresponding statistical information of the means of recording the information, in order to contextualize the user's preferences and to minimize the search time when registering the document. A search information adding device is obtained, which is characterized in that it adds search information.

Still further, according to the present invention, in a search information adding apparatus for adding a mark to a digitized document in a document management system as search key information,
A means for storing a plurality of mark attributes in advance, a means for assembling a mark by combining mark attributes, and a means for storing an assembled mark by using a mark as being composed of multidimensional attributes such as color, size, shape, etc. means for storing statistical information on the attributes of the given marks; means for storing in advance the search optimization evaluation formula; Statistical information, grant ratio instruction information.

means for calculating an optimal value using an optimization evaluation formula and extracting an appropriate mark from the mark storage means; and corresponding means for adding the extracted mark to the document and storing the statistical information. It has a means to update statistical information, and when registering a document, automatically determines the attributes of the above marks so as to minimize the time required, and assigns a mark that combines the attributes of these marks. A search information adding device is obtained, which is characterized by the following.

Furthermore, according to the present invention, there is provided a search information adding device that adds a mark to an electronic document in a document management system as search key information, in which the mark is added based on multidimensional attributes such as color, size, and shape. means for pre-storing v number of mark attributes; means for constructing a mark fr-ifl by combining the attributes; means for storing the assembled marks; Means for storing statistical information of attributes, means for pre-storing a search optimization evaluation formula, means for pre-storing instruction information of a mark attribute assignment ratio from a user, and the statistical information and assignment ratio instruction information.

means for calculating an optimum value using an optimization evaluation formula and extracting an appropriate mark from the mark storage means; displaying to a user a plurality of extracted candidate marks giving the optimum value; and a means for assigning a selected mark to the document and updating the corresponding statistical information of the means for storing statistical information, and when registering the document, the user's preference is set. reflect, grow,
A search information adding device is obtained, which is characterized in that the attributes of the mark are added so as to minimize the search time.

(Summary of the present invention) When the search information adding device of the present invention adds a mark to an input document, it uses statistical information as the usage status of the marks added so far, and statistical information for adding the mark. Based on the optimization evaluation formula and mark assignment ratio information, the optimal mark is extracted, a mark is assigned to the input document, and the statistical information of the corresponding mark is updated, so that the optimal mark is always This is a device that automatically or interactively adds marks.

Next, an example of adding a mark to a document will be shown. The criteria for evaluation is the average number of times users search for items such as marks as shown in Figure 5(a);
It is defined as the sum of the average number of searches (average number of searches S) until a desired document is obtained from a set of matching documents by specifying the conditions shown in FIG. 5(b). Now suppose that the user searches for each document with equal probability.

However, k: A, , A, l ・I in Figure 5 (a)
As shown in Ak, “mark shape” and “mark color”
Number of classification axes, such as. rj (J+==l I...
, k): Number of classification items in the j-th classification axis. For example, in classification axis A, "mark color", the classification items (1, 2, . . . , r2) are "red,""yellow," and "green."

If it is "blue", the number of classification items r is 4゜n: total number of all documents. n (i, . . . i,): The number of documents obtained when one classification item is specifically specified on each classification axis, as shown in Fig. 5. For example, when you specify "Mark Shape" as "◎", "Mark Size" as "Medium", and "Mark Color" as "Green".

The number of documents that meet the criteria. If the number of matching documents is 7, then n(i, = ◎, 12 =
Medium e'l-green) = 7°1, which is generally expressed as . That is, the first term on the right side of equation (1) corresponds to FIG. 5(a), and the second term corresponds to FIG. 5(b) K.
It corresponds to this.

At this time, it is assumed that the total number of documents n9, the number of classification axes, and the number of each classification item r3 (j=1 t 2 *..., k) have already been determined. Furthermore, in general, it is assumed that the number of documents when specifying classification items in any m (m(k)) classification axes among the classification axes is also fixed. Think about the shape of the mark, the color of the mark, and the size of the mark.
shape, size.”

The number of documents specified for each combination of "color, size" is all fixed, but "shape".

Let us consider a case where the number of documents specified by the combination K of ``color and size'' is undetermined. This is the case when n = 3 and m = 2. Therefore, in general, when the number of applicable documents up to m classification axes is known, the optimal solution based on the optimization evaluation formula when specifying the classification items of n classification axes is n-(i, ...ik+m), it becomes like a linear equation, where r; I '', r'q is 11 e
Let it be the number of classification items corresponding to ``'e lq (any q items selected from L*''+i).

From this, the optimal solution when m = 1 (it is important to give the ratio of each mark) is:

-(k-1)・n　)(3) It is.

Therefore, the following is a general description of the optimum marking procedure according to the present invention. First, in order to display or print the mark to be added, first register it in the corresponding address as code information in advance, and also record the number and type of marks added to documents registered in the past. In this situation, one new document is registered. Considering the case, first, each information of the current number of documents, the optimum evaluation formula, and the marking ratio is input into the arithmetic unit in the optimum value calculation means, and the arithmetic operation is performed.

Through this calculation, the number of documents to be added to each mark at the current time is calculated. Next, calculate the difference between the number of documents assigned in the past and determine the corresponding number of the mark that is considered to be the most appropriate to assign at this time.

The one that gives the same optimal value shall be selected from all the corresponding numbers of the corresponding mark. After that, code information for displaying or printing the mark is extracted based on a table showing the correspondence between the mark correspondence number and the address where the mark is stored.

On the other hand, the user can either automatically display only the mark corresponding to the first code information and perform the step 4, or display the marks corresponding to all the corresponding code information on the display. The grant operation is performed by selecting one of them. Finally, update the marking statistics. The above are the five basic processing procedures of the present invention. It is also conceivable that a plurality of documents are registered all at once and an appropriate mark is given to each of them, but basically < is the same as in the case of sequentially registering documents. In this case, although there is a degree of freedom in marking, it is important to note that the optimal solution is reached only after all the assumed documents have been registered. I will explain.

FIG. 1 is a block diagram showing a first embodiment of the present invention. Here, before the user actually adds search information, the following settings must be made in the mark storage means 1. Optimal evaluation formula storage means 3. and for the ratio instruction information storage means 4. The mark storage means 1 is a part in which marks to be added are registered, and may be specified in advance on the system side.Also, the mark storage means 1 is a part in which marks to be added can be registered.Also, the mark storage means 1 is a part in which marks to be added can be registered in advance. It is also possible to perform operations. Next, the optimal evaluation formula description means 3 is a part that stores an optimal evaluation formula (for example, the one described above) that minimizes the search time, and takes into account certain evaluation criteria and constraints. It is provided by the system or the user. Further, the ratio instruction information storage means 4 is a part for specifying the mark application ratio for the stored documents. This allocation ratio is set by the user. For example, four types of marks may be specified as 4:3:2:1 in descending order of frequency, or all may be specified as equal.

Further, it is assumed that the information in the statistical information storage means 2 is a part that stores statistical information of marks given in the past, and has been cleared in advance. Further, when marks are added to the input document, the number of marks added to each mark to be updated is counted and the value is stored in the statistical information storage means 2. Further, the optimum value calculation/optimum mark extraction means 5 includes the above-mentioned statistical information storage means 2.
Optimal evaluation formula storage means 3. This section calculates which mark should be added to perform optimization based on the information in the ratio instruction information storage means 4. Based on this calculated value, one mark is written 1. This is the part that extracts an appropriate mark in time for the desired mark. If a plurality of extracted optimal marks exist, any one, for example, the first extracted mark, is automatically selected. The part that actually applies the specified mark is the mark applying means 6.

Therefore, search information is added to the stored documents in the following manner, for example.

Now, the mark is either "◎", "○", or "△", and the ratio of giving this mark is, for example, ◎:○:Δ=3
: 2 : Assume that IK is set.

It is assumed that 50 documents are currently registered with optimal marks, and the following explanation will be based on the case where four new documents are to be registered. First, the mark storage means 1 stores the number of marks (number of classification items) 3. and mark "◎", "○".

"△" is stored in advance. Also, statistical information records are stored. This is the closest integer value to the grant ratio. In addition, the optimum evaluation expression means 3 stores an evaluation expression such as the above-mentioned equation (2).

Also, the mark application ratio ◎:○:Δ=3:2:1 is stored in the ratio instruction information storage means 4.

Therefore, when four new documents are input, the total number of documents becomes 54, but by using the above information in the optimum value calculation/optimum mark extraction means 5, four new documents are input. For that, mark “◎” as 2
1 item marked with "○".

It is shown that it is optimal to add "Δ" to one item. Based on this information, the information to be added to each document is automatically determined, and the mark is added by the mark adding means 6, for example, while both the document and the mark are displayed on the display. It will be done. Further, the marked information f&k is input to the statistical information storage means 2 for optimization of documents to be stored thereafter. In the case of the specific column above, there are 27 "◎" items.

There are 18 cases marked with "○" and 9 cases marked with "△". If such an assignment method is used, the burden is significantly reduced compared to the effort required to assign keywords, and the search time can be shortened.

Next, FIG. 2 shows a second embodiment of the present invention. Here, the same numbers as the means in FIG. 1 indicate means having similar functions. In addition, the user's selection means 6 is a part that displays a set of possible marks within a range that satisfies the optimum state, and selects all the marks to be added by the user from among them.

Also here, as in the first embodiment of the present invention, the settings that must be made before the user actually assigns the search information are the mark storage means 1. Optimal evaluation formula storage means 3. and for the ratio instruction information storage means 4.

Therefore, search information is added to the stored documents in the following manner, for example.

When a document is input, the optimal value calculation/optimal mark extraction means 5 receives the optimal evaluation formula from the optimal evaluation formula storage means 3, and the past marking information from the statistical information storage means 2. The mark application ratios are each transferred from the ratio instruction information storage means 4. Then, calculation is performed by the optimum value calculation/optimum mark extraction means 5,
Based on this, the optimum publication mark is extracted from the mark storage means 1. Thereafter, the AI user's selection means 7 displays a set of possible marks within the range that satisfies the optimum state, and the user selects a mark according to his/her preference from among them. In the specific example for the first example above, the new 4
For the documents, 2 have “◎” and 1 rOJ.
In this case, the system automatically assigns one piece of "Δ", but in this second embodiment, the user's selection means 7 selects a piece of ηα that satisfies the optimal condition and selects one piece as a document. The assignment is done interactively while taking into consideration the correspondence between the mark and the assignment mark. Then, at the end of the LC, the selected mark is added by the mark adding means 6.

Also, this TW report is a statistical information storage means;;
2, and provides optimization calculation information for the input document to be stored thereafter.

By using such an assignment method, draft information can be assigned according to the user's preferences while satisfying the search optimization conditions. That is, by being able to make associations to some extent during a search, it becomes easier to specify examination information and moreover, efficient examination can be performed.

Next, FIG. 1 shows a third actual example of the present invention. Here, the same numbers as the means in Figure 1 are:
This means has a similar function.

Further, the mark attribute storage means 8 is means for storing a plurality of attributes associated with a mark or a plurality of marks from different viewpoints. For example, the three attributes (classification axes) are "mark shape" and "mark size."

Think about the color of the mark. The attribute “Mark shape” is “◎”
", "○", "△" items (classification items), "Mark size" has "Large", "Medium", "Small" items, and "Mark color" has "red".

This is a case where there are items of "yellow", "green", and "affirmative". The mark attribute storage means 8 is a portion that stores in advance such classification axes, classification items, and the number of classification items corresponding to the number of classification axes.

This does not construct search information using conventional one-dimensional classification, but rather constructs search information using multidimensional classification. Here, multidimensional classification means constructing search information based on a combination of corresponding classification items of a plurality of classification axes as described above, and adding it to a stored document. The mark assembling means 9 is the part that actually assembles marks by combining such mark attributes. For example, "green" color, "big"
It is called a mark in the shape of an ``◎''. Also, the mark storage means 1 is a mark assembled by this? : This is the part that you actually remember.

The subsequent granting operations are the same as in the first embodiment. If you perform this kind of adding operation, there will be multiple clues when searching, so even if you forget what the "mark shape" was when searching for the desired document (K), you can use the "mark shape" size” and “mark color”,
Another advantage is that a search can be performed using only the "mark size." Furthermore, since the information is automatically added to minimize the search time, efficient searches are also possible.

Further, FIG. 4 shows a fourth embodiment of the present invention. ζHere, the same numbers as the means in FIGS. 1 to 3 are means having similar functions. Therefore, when such an adding device is used, search information is added according to the user's preferences while satisfying the search optimization conditions.Moreover, since there are multiple search elements, It is also possible to deal with user forgetfulness.

(Effects of the Invention) The first embodiment of the present invention has the advantage that 1 in the above formula (3)
This is the case. For example, in order to make the ratio of rOJ larger than the others for the shape of the mark ("◎", "○", "△"), the ratio of marking is specified as 1=2:1. Yes, it can be granted automatically. therefore,
Compared to the effort required to assign keywords, the burden is significantly reduced, and it has the effect of shortening the time based on the evaluation criteria of the above formula.

Furthermore, in addition to the first embodiment, the second embodiment of the present invention allows for selection of additions to be made interactively, so that assignments are made according to the user's preferences, making it possible to specify specifications that are easy to associate with. This has the effect of making the search easier.

Next, for the third embodiment of the present invention, the average number of searches when optimization is performed using the above optimization evaluation formula,
FIG. 6 shows a comparison with the average number of searches in an example without optimization. Here, we will consider a case where search information is added to 240 Dokey Men l-. This is the "mark shape"("◎").

"Q", "Δ") and "mark size" ("large".

When using "small") as search information, the sex power ratio in the form of 1 mark ◎:○:Δ=2:3:5. In addition, the mark size assignment ratio large:small=3=7 is specified in advance, and the difference between the case where optimization is performed and the case where optimization is not performed is shown. That is, the sixth
As shown in Figure (a), the assignment ratio for the size button is significant: from small = 3 near, the number of documents is 72 for "large" K out of 240 documents, and the document for "small" The number of cases has been set at 168. Similarly, as shown in Figure 6), from the ratio of marking to the shape of the mark ◎:O:△=2:3:5,
The number of documents for "◎" K is 48, the number of documents for "○" is 72, and the number of documents for "Δ" is 120. In such a situation, FIG. 6(C) shows the optimal combination of additions using the optimal evaluation formula. Further, for comparison with this, an example of a non-optimal application combination is shown in FIG. 6(d). At this time, if you simply perform a sequential search for 240 documents without adding any search information, it will take an average of 12 seconds to find the desired document.
I had to search for deer about 0 times. In addition, if non-optimal assignment as shown in FIG. 6(d) is performed, the average number of searches will be 47.1. On the other hand, Fig. 6(e)
If the optimal allocation is performed, the average number of searches will be 3.
It becomes 0. In other words, it was shown that the search time can be efficiently shortened based on the above evaluation criteria. Also,
In addition to this effect, since there are multiple pieces of search element information, for example, even if you forget information about "mark shape" K, you can use information about "mark size" to perform frugality. This has the effect that a search method can be constructed.

Furthermore, in addition to the third embodiment of the present invention, the fourth embodiment of the present invention allows selection of additions to be made interactively. This has the effect of allowing easy specification and making searches easier.

Additionally, the first to fourth embodiments of the present invention can be applied regardless of the number of documents.

Therefore, it has the advantage that it can be used to manage a large amount of documents.

The explanation above has mainly been given using examples of use based on formula (3) above, but examples of use based on formula (2), which is a film version of this, or evaluations similar to the evaluation criteria of formula (1) are also provided. Possible applications include optimizing the results. but,
They are the same in that they are added to enable efficient searches using search element information such as marks as explained above, and all are included in the present invention as long as they do not deviate from this viewpoint. .

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing a first implementation of the search information providing apparatus of the present invention. Figure 2. 3 and 4 respectively show the second section of the search information adding apparatus of the present invention. Third. It is a block diagram showing a fourth example. Furthermore, FIGS. 5(a) and 5(b) are explanatory diagrams of the average number of searches, which is one evaluation criterion for search optimization. FIG. 2 is a comparison diagram of an example using the present device and an example not using the present device for showing the same. 1... Mark storage means, 2... Statistical information storage means. 3... Optimal evaluation formula storage means. 4... Ratio instruction information storage means. 5...Optimum value calculation/optimum mark extraction means. 6... Marking means, 7... User selection means. 8...Mark attribute storage means 99...Mark assembly means. 71 Fig. O 2 Fig. 3 O 4 Fig. ρ O 5 Fig. 12 r, 12 r212 r. (b) O 6 Figure (0) (b)
(C) (d)

Claims (4)

[Claims] (1) In a search information adding device that adds marks to digitized documents in a document management system as search clue information, means for storing a plurality of marks in advance and means for storing statistical information of the added marks. and a means for pre-storing a search optimization evaluation formula; a means for pre-storing instruction information of mark assignment ratio from the user; and using the statistical information, assignment ratio instruction display information, and optimization evaluation formula, means for calculating an optimum value and extracting an appropriate mark from the mark storage means; and means for adding the extracted mark to the document and updating corresponding statistical information in the means for storing statistical information. . A search information adding device characterized in that when a document is registered, the mark is automatically determined and added so as to minimize the search time. (2) In a search information adding device that adds marks to digitized documents in a document management system as search clue information, means for storing a plurality of marks in advance and means for storing statistical information of the added marks. , a means for pre-storing a search optimization evaluation formula, a means for pre-storing instruction information of the mark assignment ratio from the user, and an optimum value using the statistical information, the assignment ratio instruction information, and the optimization evaluation formula. means for calculating the mark and extracting an appropriate mark from the mark storage means; means for displaying a plurality of extracted candidate marks giving the optimum value to the user and having the user select one; It has a means for adding a mark to a previous document and updating corresponding statistical information in the means for storing said statistical information, so that when registering a document, the user's preferences are reflected and the search time is minimized. A search information adding device characterized in that it adds the mark as described above. (3) In a search information adding device that adds marks to digitized documents in a document management system as search clue information, the marks can be changed by color, size, etc.
a means for storing a plurality of mark attributes in advance, a means for assembling a mark by combining the mark attributes, and a means for storing the assembled mark; means for storing statistical information on the attributes of assigned marks; means for pre-storing a search optimization evaluation formula; means for pre-storing instruction information for mark attribute assignment ratios from users; and the statistical information and assignment ratios. means for calculating an optimal value using instruction information and an optimization evaluation formula and extracting an appropriate mark from the mark storage means; and means for adding the extracted mark to the document and storing the statistical information. When registering a document, it automatically determines the attributes of the above marks and creates a mark that combines the attributes of these marks, so as to minimize the search time. A device for adding search information, characterized in that it adds search information. (4) In a search information adding device that adds marks to digitized documents in a document management system as search clue information, marks can be added by color, size, etc.
a means for storing a plurality of mark attributes in advance, a means for assembling a mark by combining the mark attributes, and a means for storing the assembled mark; means for storing statistical information on the attributes of assigned marks; means for pre-storing a search optimization evaluation formula; means for pre-storing instruction information for mark attribute assignment ratios from users; and the statistical information and assignment ratios. Means for calculating an optimal value using instruction information and an optimization evaluation formula and extracting an appropriate mark from the mark storage means, and displaying to a user a plurality of extracted candidate marks that give the optimal value. means for causing the user to select one; and means for adding a selected mark to the document and updating corresponding statistical information in the means for storing statistical information; A search information adding device characterized in that the attributes of the mark are added so as to reflect the preferences of users and to minimize the search time.