JPS62175863A - Slip processor - Google Patents

Abstract

PURPOSE:To check whether the input data is made coincident or not with the types of items and the format by obtaining automatically the type and the format of data on each item with use of an analysis means together with a slip data memory means and a memory means for dictionary information on those items. CONSTITUTION:Here the names of five persons in charge, TANAKA, OGAWA TARO, SUZUKI, Suzuki and FURUKAWA are written on five slips respectively and the dictionaries of each person name are stored in a dictionary memory device 6 as shown by figures (a)-(e). An analysis means 3 leaves the attribute that holds a majority with a certain value among those attributes based on the data on these five dictionaries (1). An OR connection is secured for the available value with said remaining attribute (2) and the AND connection is secured among attributes (3). Then >=3 elements having the same host words if obtained in the obtained conditional equation are replaced with the host words respectively (4). Here the numerical data such as the amount of money, date, etc., are treated especially (5). Thus the type and the department remain as attributes in the process (1) and the type (person name) and the department (department A or B) are obtained in the process (3).

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a form processing device that analyzes form data entered in each item and learns its type and format.

[Technical background of the invention and its problems]

With the development of information processing technology, it has become possible for computers to handle forms (slips, request forms, etc.) that were normally handled in office work, and t0 electronic form processing devices display form formats (form data) on the display. It is a device that allows general users, even those without computer knowledge, to easily enter, change, edit, and tabulate data.

By the way, when I try to enter data into a form, I sometimes don't know what format to use. Also, when I look at the item ``in charge'', I often don't know whether it means ``name of person in charge'' or ``district in charge.'' Normally, such cases are dealt with by examining several examples of the form data that have already been filled out. If the form is not electronic, there is always an example of how to fill it out, and I look at it while filling out the form. However, in the case of electronic forms, it is necessary to operate the keyboard etc. to display the information on the display, making the most basic and important operation of entering data difficult for people who are not accustomed to using the device. There was a problem. In order to solve this problem, you can design the form of the form on a computer, and if you fill in this item in such and such a format, you may be able to create a 7'h7t opening. However, this makes designing the form form very troublesome. In reality, the problem was that there were only nine people who could create or not create friends and animals, and it would not become a monument to unification.

[Purpose of the invention]

The purpose of the present invention is to check the data entered by the user when the user enters the data, and to notify the user of inappropriate data, thereby facilitating the data entry, as well as checking the data. The purpose is to provide a good form processing device without spending extra effort on the details.

[Summary of the invention]

The present invention automatically determines the type and format of data for each item by means of an analysis means from a form data storage means that stores form data and a dictionary storage device that stores dictionary information regarding the items. This is a check to see if the data matches the type and format of the item.

〔Effect of the invention〕

According to the present invention, the user is allowed to fill in one type and four types without worrying about it. This is because if it is incorrect, it will be checked by the checking mechanism. Furthermore, this checking mechanism is highly reliable because it analyzes what kind of description is allowed based on the actual data input so far. Furthermore, since the information on 7.1 part type and *formula is automatically obtained, it does not require any extra effort on anyone's part. From the above points, the practicality of the present invention is enormous.

[Embodiments of the invention]

Hereinafter, one embodiment of the present invention will be described with reference to the drawings. FIG. 1 is an overall configuration diagram of an embodiment of the present invention. The form data storage device 4 stores form data input via the keyboard 1 and displayed via the display 2. The dictionary storage device 6 stores dictionary information S corresponding to each item of form data. Analysis means 3 is
The type and format of each item data is determined from the dictionary storage device and the form data storage device, and is written in a fixed format. On the other hand, the checking mechanism 5 compares the input data with the analysis results previously performed, and if it is correct, it does nothing, but if it is not, it notifies the user. The analysis may be performed at the time the checking mechanism operates.

This will be explained in detail by taking the form shown in FIG. 2 and the processing flow diagram of the analysis means shown in FIG. 7 as examples. Consider determining the type and format of the data for the item “in charge” in Figure 2 (Section 7).
Figure 71.72). Currently, the person in charge of each of the five 1-kuma slips is written as Nichichi, Koyotabe, Suzuki, Suzuki, and Furukawa, and the dictionary for each person's name is stored in the dictionary storage device 6 as shown in Figure 3 (al to t).
Suppose that it is stored as el. Analysis means 3 uses the following method to analyze data from these five dictionary data.
Find the ffi types (Fig. 7 73).

(1) Leave the attributes that have a majority with a certain value in the @-character.

(2) For the remaining attributes, OR-combine the possible values.

(3) Make an AND connection between attributes.

(4) There are 3 elements with the same hypernym in the obtained conditional expression.
If there are more than one, replace them with hypernyms.

(5) Numerical data (full amount on a single date, etc.) will be treated specially.

According to (1), type and affiliation remain as attributes. Each person's name and Ail represent the majority. C2).

(3) gives the following conditional expression.

Type - person name and affiliation - (A section or 8 section) (4) is not applied here, but for example, in the above formula, affiliation = (A section or 8 section)
Afi as Narano in Section B or Section C].

Suppose that all the hypernyms of sections B6'A and C are part 8.

Then, according to (4), this condition becomes <=Sm. In this way, 4 of the data! i'JA is found (Fig. 7 74
).

Next, find the format of the data. The first thing to consider in the data format is whether it is a Kanji code, a numerical value, or a character code. Kanji codes are further divided into 3 (1M): hiragana, katakana, and kanji.

The numerical value is divided into integers and decimals, and then divided into h3 digits. Find the data format using the following method.

(1) Distinguish whether it is a kanji code, a numerical code, or a character code.

(2) If it is a kanji code, distinguish between hiragana, katakana, and kanji.

(3) If the number is straight, distinguish whether it is an integer or a decimal and find the number of digits.

(4) If the item is date-related, decide the date format type.

+5) For (1) to (4), examine the individual data and select the most common format for that item.

Applying the example in FIG. 3 to (1) to (5), the format = kanji code and kanji. As another example, consider the "date" item.The data is 850407.841019,8404.600527
Assume that it is 850,630°.

When examining individual data, 850407.841019.850630...
・・・・ Typee (2 digits, 2 digits for October, 2 digits for 10th) 8404 ・・・・・・・・・・・・
・Type ■ (2 digits of the Western calendar, 2 digits of October) 600527 ・・・・・・・・・
... Since it can be classified as Type ■ (Showa 2 digits, October 2 digits, 10th 2 digits), the majority vote in (5) will result in type. In the end, the format is: date = integer and 6 digits and typeface (Figure 7, 75).

As described above, the actual data is analyzed and the most probable conditions are defined (76 in Figure 7). The determined types and conditions of the qI formula are expressed in the format shown in FIG. 4 ((1) fb). Types are represented by attribute name and value pairs. The place where you are in charge of looking for the future @ - Figure 5 (a) (b)
become that way. The format for type 1 is kanji code, numerical value,
Type 2 is a distinction between character codes, hiragana, katakana, and kanji, or a distinction between integers and decimals. The number of digits is the number of digits in the case of numerical values, and the date type is the type in the case of date-related items.

Next, the checking mechanism will be explained using the processing flow diagram of FIG.

The types and formats of the nine items obtained by analysis means 3 are determined in the format shown in Figure 4. The input data is processed as either dictionary data obtained from the dictionary storage device 6 or not in the dictionary (FIG. 8, 81 and 82). Then, it is checked whether conditions such as No. 51g are met (FIG. 8, 83).

For example, if "Sawara" is entered in the person in charge column, the data shown in FIG. 6 is obtained from the dictionary storage device 6 and is assumed to be 9.

As for the type of attribute name, it is determined that "Sawara" is incorrect because either the person's name matches or the affiliation does not match.Alternatively, if "°600401" is input as a date, it is first judged as a numerical value and becomes a 6-digit integer. Then, the date type is checked and it becomes type ■, which is also determined to be incorrect (
Figure 8 84). In this way, the checking mechanism checks whether the input data satisfies the conditions based on the analysis results and outputs a message indicating "incorrect" if the data does not satisfy the conditions (FIG. 8, 85).

It should be noted that the present invention is not limited to the above embodiments. First, there are various considerations regarding when to perform the analysis using the first analysis method. It can be done every time there is input data, it can be done at regular intervals, or it can be done at a certain time. The first two methods are repeated %(1') times, so the analysis results change.As the analysis and learning process progresses, the conditions change little by little.Whether or not these conditions approach the correct conditions depends on the statistical properties of the input data and the learning process. It cannot be generalized because it varies greatly depending on the method of learning.There are many different methods of learning.

For example, in the example, an attribute that has the same value for a majority of cases is adopted, but an attribute that does not have more than 4 types of values is adopted, or the majority is changed to 30% or 40%, or the correlation is determined by looking at the relationship between attributes. For example, only certain attributes are considered. Regarding the format, dictionary data may also be used to define the format in more detail.

For example, add data indicating whether it is just the last name or the first name in the data for the person in charge. For dates, it is also possible to limit the values to this year and last year. Whether abbreviations are allowed or not
How to define various formats that are related to the meaning of the item, such as whether numerical data has units or not, and from what level (prefecture, city, town, village, etc.) should be entered in the case of an address entry item. There is r1.

In short, the present invention can be modified in various ways without departing from the spirit thereof.

[Brief explanation of drawings]

Figure 1 (G, fruit of the present invention'+1 Outline of Shio J
Figure 2 shows the format of the ratio form used as columns, Figure 3 shows the columns of the dictionary, Figure 4 shows the format for describing the analysis results, and Figure 5 shows an example of how to write the analysis results. 6 is a diagram showing an example of a dictionary of input data Sahara°, FIG. 7 is a processing flow diagram of the analysis means, and FIG. 8 is a processing flow diagram of the checking mechanism. l...Keyboard, 2...Display, 3...
Analysis means, 4-- form data storage 47m, 5-- checking mechanism, 6-- dictionary data ife storage device. Agent: Patent attorney Chika Nori, @ Yudo Kikuo Takehana Figure 2 Figure 6 (α) (6) CC)
Figure 7

Claims (1)

[Claims] a form data storage means for storing form data consisting of a plurality of items; a dictionary storage means for storing dictionary information regarding the items of the form data; an analysis means for determining the type and format of data to be filled in using dictionary information;
A form processing device comprising: checking means for checking the data entered in the form data according to the type and format determined by the analysis means.