JP2002334045A - E-mail classification method, its implementation device and its processing program - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To provide a technique capable of automatically weighting a sent mail or a received mail reflecting the user's interest. In an e-mail classification method for classifying e-mails according to their importance, a step of extracting a set of words from an e-mail text of an e-mail whose importance is to be registered, Assigning a value representing the importance to the set of words, registering the extracted word set and its importance in the determination reference data, and Extracting a set of words from the text, reading the importance of the extracted set of words from the criterion data, and classifying the e-mail according to the read value of the importance. Have

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an e-mail classifying apparatus for automatically classifying e-mails, and more particularly to an e-mail classifying apparatus for classifying e-mails that sufficiently reflects the interest of a user. And effective technology.

[0002]

2. Description of the Related Art Conventionally, the classification of received e-mail
There were two formulas. One is based on the attributes of the mail, such as the title and sender of the mail, and the other is based on the body of the mail.

There are two methods based on the body of the mail, a classification method according to the contents of the mail and a classification method according to the user's interest. The former method can be realized by a method in which words in an e-mail are referred to in a vocabulary classification dictionary in which words are classified for each field, and a word in which field appears frequently is used as a criterion. The latter could be realized by using data in which the user himself / herself registered words of interest as classification criteria instead of the general vocabulary classification dictionary used in the former method.

A method has been proposed in which a word of interest registered by a user is used as a reference for classification (Japanese Patent Laid-Open No. 9-305505). A method in which a keyword is automatically extracted from a received mail determined to be unnecessary by the user and an unnecessary mail is determined based on the keyword (JP-A-11-15756).
Has also been proposed.

[0005]

In the above-mentioned prior art, there is a problem that a user must first register a word on the assumption of an object of interest. Also, even in the method of automatically extracting keywords, the keyword is extracted only for received mails, and only judgments for the necessity / unnecessity are described, and classifications that sufficiently reflect the interests of users cannot be performed. There was such a problem.

An object of the present invention is to solve the above-mentioned problem and to provide a technique capable of automatically weighting a sent mail or a received mail reflecting the user's interest.

[0007]

SUMMARY OF THE INVENTION According to the present invention, there is provided an e-mail classification apparatus for classifying e-mails according to their importance, which classifies e-mails according to the importance of a set of words extracted from the mail body. Is what you do.

[0008] In the electronic mail classification device of the present invention, a set of words is extracted from the mail body of the sent mail or the received mail whose importance is to be registered, and the extracted set of words is received from the user. After assigning a value representing the degree of importance of the designated degree of interest, the extracted word set and the degree of importance are registered in the criterion data.

Next, a group of words is extracted from the mail body of the electronic mail whose importance is to be determined, and the importance of the extracted group of words is read from the determination criterion data. The importance of each pair of words extracted from the text is added to determine the importance of the email,
The degree of importance of the electronic mail is displayed according to the value of the obtained importance.

As described above, in the present invention, there is no need for the user himself to create a weighting standard for the electronic mail.
The user automatically creates a criterion by simply specifying whether he or she is interested in outgoing mail or received mail, and based on this criterion, determines whether the importance of the mail is high or low, By displaying numerical values, e-mails can be easily arranged and sorted. Also, the automatically created weighting criterion is used to make more accurate judgments,
The user may arbitrarily register and delete words.

As described above, according to the e-mail classification device of the present invention, e-mails are classified according to the importance of a set of words extracted from the main body of the mail. It is possible to automatically perform weighting reflecting the interest of the person.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An electronic mail classification apparatus according to an embodiment for classifying electronic mails according to their importance will be described below.

FIG. 1 is a diagram showing a configuration of an electronic mail transmission / reception system using an electronic mail automatic classification function according to this embodiment. As shown in FIG. 1, the e-mail transmission / reception system of the present embodiment includes a mail importance registration unit 32 and a mail importance determination unit 33.

The mail importance registering unit 32 extracts a set of words from the mail body of the electronic mail whose importance is to be registered, and assigns a value indicating the importance to the extracted word set. Then, the processing unit registers the extracted word set and its importance in the determination reference data.

The mail importance judging section 33 extracts a set of words from the mail text of the electronic mail whose importance is to be judged, and reads the importance of the extracted word set from the judgment reference data. A processing unit for classifying the electronic mail according to the read importance value and displaying the degree of importance.

A program for causing the electronic mail transmission / reception system to function as the mail importance registration section 32 and the mail importance determination section 33 is recorded on a recording medium such as a CD-ROM, stored on a magnetic disk or the like, and then stored in a memory. It shall be loaded and executed. The recording medium for recording the program may be a recording medium other than the CD-ROM. The program may be installed from the recording medium to the information processing apparatus and used, or the recording medium may be accessed through a network to use the program.

In this embodiment, the server 10 and the client 30 are allocated on the network 20.

The mail server 11 in the server 10 manages mail transmission and reception with the client 30 via the network 20. The client 30 has a mail communication unit 3
1. Mail importance registration unit 32, mail importance determination unit 3
3, a mail storage unit 34, and determination reference data 35 are assigned. The user environment is constituted by the display device 40 for displaying the contents of the mail and the like, and the input device 50 for performing an operation.

The mail information transmitted and received via the mail communication unit 31 is stored in the mail storage unit 34. An arbitrary mail in the mail storage unit 34 is input to the input device 50.
When the user performs an operation for setting the importance, the e-mail importance registration unit 32 creates judgment reference data from the specified e-mail, and registers the result in the judgment reference data 35. At this time, the operation for setting the importance is performed by the mail storage unit 3.
4 can be executed for one or more mails, and can be executed at a plurality of timings, such as when reading or deleting a mail.

Next, when the user performs an operation of judging the importance of an arbitrary mail in the mail storage unit 34 by using the input device 50, the mail importance judging unit 33 outputs the judgment reference data to the specified mail. The importance is determined with reference to 35, and the result is displayed on the display device 40.
At this time, the operation of the importance determination is executable for one or more mails in the mail storage unit 34.

FIG. 2 shows the mail importance registering section 3 of the present embodiment.
6 is a flowchart illustrating a processing procedure of No. 2; There are several methods for expressing the characteristics of a document by numerical values. Here, a case where the method is realized by using a method of expressing a document by a vector will be described as an example. The method of expressing a document by a vector is a method in which, for one word in a sentence, another word related to the word is set as one set, and these words indicate characteristics of the document. The mail importance registering unit 32 creates determination reference data from the transmitted / received mail by the following processing.

First, in step 201, the mail importance registration section 32 causes the word extraction processing section 60 to extract words from the mail body. The method of extracting words includes a method of extracting words by referring to dictionary data, a method of extracting characters using characters such as kanji and hiragana in a sentence as a mark, and the like, but the method is not particularly limited here.

Next, in step 202, the inter-word relationship determination processing section 61 searches for a word that is related to one word. Here, the relation between words means a relation based on Japanese grammar such as a subject, a verb, and a dependency. This can be realized by using a general Japanese analysis method based on the meaning of particles, words, and the like.

Next, in step 203, the degree of interest specified by the user with respect to the text of the mail is quantified by the importance setting processing unit 62 for the set of words extracted by the inter-word relation determination processing unit 61. , As a value indicating importance. Regarding the numerical value of the degree of interest, use a value specified directly by the user as a numerical value, or a value specified by an ambiguous expression such as high / low, when "high" is +10, when "low" Can be replaced with a predetermined numerical value such as -10. It is also possible to give each word set a difference in importance by multiplying the numerical value indicating the importance by the number of times this word set appears in the sentence.

Next, at step 204, the existing data comparison / registration processing section 63 compares the data created by the importance setting processing section 62 with the existing judgment criterion data 35, and if the same word pair does not exist, Goes to step 205, and if there is, goes to step 206.

In step 205, the importance setting processing section 6
In step 206, the data created in step 2 is added to the criterion data 35. On the other hand, in step 206, the result of adding or subtracting the respective degrees of importance is registered again in the criterion data 35 to reflect a change in interest. In the present embodiment, a process equivalent to subtraction is performed by setting a negative value to the importance.

The determination reference data 35 can be automatically generated by the above-described processing procedure. However, the user can directly register / delete the data or change the numerical value indicating the importance. It is also possible to create a standard that more closely matches the user's intention.

FIG. 3 shows the mail importance judging unit 3 of the present embodiment.
13 is a flowchart illustrating a processing procedure of No. 3; The mail importance determining unit 33 determines the importance of the transmitted / received mail by referring to the criterion data 35 generated by the mail importance registering unit 32.

First, in step 301, the mail importance determining section 33 causes the word extraction processing section 70 to extract words from the specified mail body. This processing is performed by the word extraction processing unit 60.
This is the same processing as. Next, in step 302, the inter-word relationship determination processing unit 71 forms a set of related words.
This processing is the same as the processing performed by the inter-word relationship determination processing unit 61.

Next, in step 303, the importance reference processing unit 72 searches the existing determination criterion data 35 using the word set created in the inter-word relationship determination processing unit 71 as a search key. The contents of the search result are examined, and if the word set created by the inter-word relationship determination processing unit 71 is registered in the existing determination reference data 35, the process proceeds to step 305. If not, the process proceeds to step 306. move on.

In step 305, the criterion data 35
By setting the numerical value of the importance in the above, the numerical value of the importance is given to the set of words created by the inter-word relationship determination processing unit 71. On the other hand, in step 306, a numerical value indicating no importance is set for the word set.

In step 307, the mail weight calculation processing unit 73 adds the importance set to all the word sets created by the inter-word relation determination processing unit 71 to determine the importance for the mail. Calculate and classify the email, and according to the calculated importance value, high / medium /
A display indicating the importance of the mail such as low is displayed.

FIG. 4 shows the mail importance registering section 3 of the present embodiment.
FIG. 11 is a diagram showing an example of an implementation screen of mail importance registration using No. 2; The mail transmission / reception entire screen 80 includes buttons for operation and a mail list screen. The operation buttons include a new mail creation button 81, a mail transmission / reception button 82, a mail deletion button 83, a mail importance determination button 84, and a mail importance setting button 85. The mail list screen includes a received mail list 86 and a transmitted mail list 87.

Here, the operation in the case of performing the importance registration operation for one received mail will be described. First, a target mail is selected from the received mail list 86. In the figure, the mail with the “X” mark in “Select” is the selected mail.
Next, one of high / low of the mail importance setting button 85 is pressed according to the interest in the mail. Here, it is assumed that "high" is pressed because the user is interested in the contents of the mail. In the example, the degree of interest is selected from either high or low, but more detailed designation using numerical values, graphs, or the like is also possible.

FIG. 5 is a diagram showing an outline of the process of registering the mail importance of FIG. 4 of the present embodiment. The method of setting the numerical value of importance and the method of calculation described here are examples of a realization method.

First, a word is extracted from the mail body 90 in which the user has specified interest “high”, and a word extraction result 91 is obtained. Next, a set of related words is set from the word extraction result 91, and the set of words 92 to which the appearance frequency in the sentence is added to the set.
Generate "Email / infection" is 2
Since it appears twice, the frequency is 2. The frequency is 1 since the other sets appear only once.

Next, the degree of interest specified by the user is reflected. Here, since the specification of the user is not specified as a numerical value, the value in the case of interest "high" is set as +10. Further, here, it is assumed that a process of multiplying the importance by the appearance frequency is performed, and the frequency is multiplied by the frequency of the word set 92, and importance setting data 93 in which the result is assigned to the word set is generated.

Next, the importance setting data 93 is compared with the existing judgment criterion data 94, and if there is the same word set, the importance is added or subtracted, and the completion judgment criterion reflecting the change in interest is reflected. The data 95 is generated. In the example, the combination of the words “virus / infection” matches, and the importance is −10 in the existing determination criterion data 94 and +10 in the importance setting data 93 generated from the mail 90. The severity of "infection" was set to 0.

FIG. 6 shows the mail importance judging section 3 of the present embodiment.
FIG. 13 is a diagram showing an example of an implementation screen for e-mail importance determination using No. 3. The entire mail transmission / reception screen 100 is the same screen as that of FIG.

First, in the outgoing mail list 102, a target mail is selected. In the figure, the target mail is one in which "selection" is marked with a cross. Next, when the importance determination button 101 is pressed in order to determine the importance of the specified outgoing mail, for example, if the importance is greater than 100, “high”; If it is less, it is determined as “low”. Here, as a result of the determination, a display such as high / middle / low is displayed, but it is also possible to display in more detail by a numerical value, a graph, or the like.

FIG. 7 is a diagram showing an outline of the processing for judging the importance of mail in FIG. 6 of the present embodiment. Even if the user specifies a plurality of mails, the determination process is performed for each mail. Here, word extraction is first performed on one specified mail body 110, and a word extraction result 111 is obtained.

Next, a related word set 112 is obtained, and a word set 112 to which the appearance frequency in the sentence is added is obtained. This is compared with the existing determination criterion data 113, and when the same word set exists, the importance setting data 114 is generated by referring to the numerical value of the importance set in the existing determination criterion data 113. In this example, the set of the words “vaccine / detection” and “infection / detection” does not exist in the existing criterion data 113. “Virus / infection” is present, but its importance is 0. For this reason, only the importance of “e-mail / infection” existing in the existing determination reference data 113 +20 is used, and the result obtained by multiplying the frequency by 1 is set in the importance setting data 114.

The result obtained by adding all the degrees of importance of the degree of importance setting data 114 is defined as the degree of importance 115. In this example, the degree of importance is +2.
It became 0. By using the value of the importance 115, it is possible to indicate the level of importance of the mail.

As described above, according to the present embodiment, it is possible to know the importance according to the user's interest without referring to the contents of the transmitted and received e-mails, so that the mails can be easily sorted. It becomes possible. In addition, by creating a criterion for determining the importance from the transmitted / received mail, it is possible to cope with a change in the importance due to a change in interest.

Further, in the present embodiment, not only the importance of the received mail but also the importance of the transmitted mail can be known. Deletion depending on the degree, etc.
Useful for organizing and sorting outgoing mail. The importance determined in the transmitted mail may be sent to the destination together with the transmitted mail, and may be used as the importance on the receiving side.

As described above, according to the e-mail classification device of the present embodiment, e-mails are classified according to the importance of a set of words extracted from the main body of the mail. It is possible to automatically perform weighting reflecting the interest of the user.

[0047]

According to the present invention, e-mails are classified according to the importance of a set of words extracted from the main body of the mail. Therefore, weighting that reflects the user's interest in outgoing mail or received mail is performed. Can be performed automatically.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of an e-mail transmission / reception system using an automatic e-mail classification function according to an embodiment.

FIG. 2 is a flowchart showing a processing procedure of a mail importance registration unit 32 of the embodiment.

FIG. 3 is a flowchart showing a processing procedure of a mail importance determining unit 33 of the embodiment.

FIG. 4 is a diagram illustrating an example of a screen for performing mail importance registration using the mail importance registration unit 32 according to the embodiment.

FIG. 5 is a diagram showing an outline of processing for registering the mail importance of FIG. 4 in the embodiment.

FIG. 6 is a diagram illustrating an example of a screen for performing a mail importance determination using the mail importance determination unit 33 according to the embodiment.

FIG. 7 is a diagram showing an outline of a process of judging the mail importance of FIG. 6 according to the embodiment;

[Explanation of symbols]

10 server, 11 mail server, 20 network, 30 client, 31 mail communication unit, 34
E-mail storage unit, 35... Determination reference data, 40... Display device, 50... Input device, 60.
Existing data comparison / registration processing unit, 70 ... word extraction processing unit,
71: Inter-word relationship determination processing unit, 72: Importance reference processing unit, 73: E-mail weight calculation processing unit, 32: E-mail importance registration unit, 33: E-mail importance determination unit, 80: E-mail transmission / reception entire screen, 81 ... New mail button, 82 ...
Email send / receive button, 83 ... Email delete button, 84 ...
Mail importance judgment button, 85: Mail importance setting button, 86: Received mail list, 87: Sent mail list, 9
0: mail text, 91: word extraction result, 92: word set,
93 ... importance setting data, 94 ... existing judgment reference data,
95: Completion judgment reference data, 100: Mail transmission / reception entire screen, 101: Importance judgment button, 102: Outgoing mail list, 110: Mail text, 111: Word extraction result, 1
12 ... word set, 113 ... existing criterion data, 114 ...
Importance setting data, 115 ... Importance.

Claims (5)

[Claims] 1. An e-mail classification method for classifying e-mails according to their importance, a step of extracting a set of words from an e-mail body of the e-mail whose importance is to be registered, Assigning a value indicating importance to the set of words;
Registering the extracted word set and its importance in the criterion data; extracting the word set from the mail body of the electronic mail whose importance is to be determined; A step of reading the importance of the set from the determination criterion data, and a step of classifying the electronic mail according to the read value of the importance. 2. The method according to claim 1, wherein a degree of importance input by a user is set in the criterion data. 3. The method according to claim 1, wherein the degree of importance of the electronic mail is displayed according to the read value of the importance. 4. An electronic mail classification device for classifying electronic mails according to their importance, extracting a set of words from the mail body of the electronic mail whose importance is to be registered, and A mail importance registering unit for assigning a value indicating importance to the group and registering the extracted word group and its importance in the determination reference data; and A mail that extracts a set of words from a mail body, reads the importance of the extracted set of words from the determination criterion data, and displays the level of importance of the electronic mail according to the read value of the importance. An e-mail classification device, comprising: an importance determination unit. 5. A program for causing a computer to function as an e-mail classification device for classifying e-mails according to their importance, wherein a set of words is extracted from the e-mail body of the e-mail whose importance is to be registered. A mail importance registering unit that assigns a value indicating importance to the extracted word set, and registers the extracted word set and its importance in determination reference data; A set of words is extracted from the mail body of the target e-mail, the importance of the extracted set of words is read from the determination criterion data, and the value of the e-mail is read in accordance with the read value of the importance. A program for causing a computer to function as a mail importance determining unit that displays the level of importance.