WO2020175371A1

WO2020175371A1 - Email thread extraction device, email thread extraction method, and computer program

Info

Publication number: WO2020175371A1
Application number: PCT/JP2020/007065
Authority: WO
Inventors: 宏一千葉; 孝治　吉春
Original assignee: 株式会社エー・アンド・ビー・コンピュータ
Priority date: 2019-02-25
Filing date: 2020-02-21
Publication date: 2020-09-03
Also published as: JPWO2020175371A1; JP6905237B2

Abstract

Provided is an email thread extraction device that is capable of using a learned model that uses a neural network to extract an email exchange as a thread completely and suitably. The email thread extraction device comprises: an email data storage unit for storing a plurality of items of email data; an origin email data acquisition unit for acquiring, from the email data storage unit, source email data serving as an origin of an email thread; a reply candidate data generation unit for using a learned model, wherein a correspondence relationship between email data and reply email data corresponding thereto has been learned, to generate reply candidate data corresponding to the origin email data; a reply email data acquisition unit for acquiring, from the email data storage unit, reply email data similar to the reply candidate data generated by the reply candidate data generation unit; and an email thread generation unit for generating an email thread in which the origin email data and the reply email data are combined.

Description

\\0 2020/175371 1 ?<:17 2020/007065 Clarification

Title of invention:

Mail thread extracting device, mail thread extracting method, and computer program

Technical field

The present invention relates to a mail thread extracting device for collectively extracting mail exchanges (mail threads).

Background technology

[0002] Conventionally, various techniques have been proposed for managing electronic mail for each common topic. For example, Patent Document 1 (Patent No. 5 919 449) discloses a mail data classification device for classifying electronic mail data, such as a threaded display of electronic mail. In this mail data classification device, the words contained in the email and the response phrase that is the reply word to the word are stored in association with each other, and the keyword relating to the word contained in the email to be classified is searched for, and the response phrase is searched. E-mails containing are extracted, and these are combined and displayed in the same group.

[0003]Further, a technique of collectively extracting e-mail exchanges using "title", "destination", and the like, which are information in the header part of e-mail, has been widely used.

Prior art documents

Patent literature

[0004] Patent Document 1: Patent No. 5 1 9 9 4 4 9

Summary of the invention

Problems to be Solved by the Invention

However, thread extraction using keywords has a problem that it is difficult to obtain appropriate extraction results unless the keywords are set properly. Also, " 〇 2020/175371 2 卩 (: 170? 2020 /007065

With thread extraction using "Title" and "Destination", even if the email has a conspicuous content, if the "Title" and "Destination" are different, it will not be properly organized as a thread. There is also. Furthermore, recently, the use of artificial intelligence (8) is becoming more practical, and it is possible to extract email threads according to the content of the email using a trained model using a neural network. It is assumed as an application field of I.

[0006] The present invention is a mail thread extracting device and a mail thread extracting method capable of appropriately extracting mail exchanges as mail threads using a trained model using a neural network. And to provide computer programs and the like.

Means for solving the problem

In order to achieve the above object, a mail thread extracting device of the present invention is a mail data storage unit for storing a plurality of electronic mail data,

A starting point mail data acquisition unit that acquires starting point mail data that is the starting point of the mail thread from the mail data storage unit,

A response candidate data generation unit that generates response candidate data corresponding to the origin mail data by using a learned model that has learned the correspondence between the mail data and the corresponding response mail data,

A response mail data acquisition unit that acquires response mail data similar to the response candidate data generated by the response candidate data generation unit from the mail data storage unit;

A mail thread generation unit that generates a mail thread by combining the origin mail data and the response mail data is provided.

Effect of the invention

[0008] According to the present invention, a mail thread extracting device and a mail thread extracting device capable of appropriately extracting e-mail exchanges as mail threads without omission by using a trained model using a neural network. It is possible to provide a code extraction method and a computer program. 〇 2020/175371 3 卩(: 170? 2020/007065 Brief explanation of the drawing

[0009] [Fig. 1] Fig. 1 is a block diagram showing a schematic configuration of a mail thread extracting device according to a first embodiment of the present invention.

[Fig. 2] Fig. 2 is a diagram showing an example of making a document into a vector.

[FIG. 3] A diagram showing an example of mail data stored in a mail data storage unit. 4 is a diagram showing an ^_ example of a user in tough Ace the origin mail data and the search period in the first embodiment to specify the present invention.

FIG. 5 is a diagram showing an example of a mail thread according to the first embodiment of the present invention. FIG. 6 is a diagram showing an example of exchange of electronic mail.

FIG. 7 is a diagram showing an example of a set of mail data and response mail data used as teacher data.

[Fig. 8] Fig. 8 is a diagram showing an example of a neural network for learning the correspondence relationship of electronic mail.

[FIG. 9] A flow chart showing an example of a processing flow of mail thread extraction according to the first embodiment of the present invention.

FIG. 10 is a block diagram showing a schematic configuration of a mail thread extracting device according to a second embodiment of the present invention.

[Fig. 11] Fig. 11 is a flow chart showing an example of a processing flow of mail thread extraction according to the second embodiment of the present invention.

FIG. 12 is a diagram showing an example of a user interface for designating a starting point email data and a search period in the second embodiment of the present invention.

FIG. 13 is a diagram showing an example of a mail thread according to the second embodiment of the present invention.

[FIG. 14] A diagram showing an example of a neural network for learning a similar relationship of response mail data according to a third embodiment of the present invention.

MODE FOR CARRYING OUT THE INVENTION

[0010] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the drawings, the same or corresponding parts will be denoted by the same reference symbols and description thereof will not be repeated. 〇 2020/175371 4 卩 (: 170? 2020 /007065

[001 1] [First embodiment]

FIG. 1 is a block diagram showing a schematic configuration of a mail thread extraction system 100 according to the first embodiment of the present invention. The mail thread extraction system 100 extracts the mail data that is the starting point and the mail data that responds to it according to the specified search conditions, and creates a mail thread by combining the extracted mail data. However, the mail thread extraction system 100, like the conventional mail thread extraction system, simply extracts the response mail data depending on whether the text data of the subject or the body contains a predetermined keyword. The response mail data is extracted and the mail thread is generated by using the trained model generated based on a large amount of training data, instead of generating the mail thread.

As shown in FIG. 1, the mail thread extraction system 100 includes an extraction device 1 and a learning device 2. The extraction device 1 and the learning device 2 do not need to be always connected. For example, the learner 2 can be configured as a cloud system.

The extraction device 1 includes a mail data analysis unit 11, a mail data storage unit 12, an origin mail data acquisition unit 13, a response candidate data generation unit 14, a response mail data acquisition unit 15 and a mail thread. It includes a code generation unit 16.

[0014] The mail data analysis unit 11 morphologically analyzes the mail data of the electronic mail and then converts it into a vector. The vectorization of mail data is generally performed by a dictionary that is composed of the morphemes of document data and their mouths. Figure 2 shows the dictionary conversion of document data. Document data is made into a dictionary by morphologically analyzing the underlying mail data and giving them a unique clue. Each I mouth represents the dimension of the analytic space composed of a dictionary. Vectorization of mail data is performed by morphological analysis of the mail data to be vectorized, matching with the morphemes in the dictionary, and expressing the number of hits for each mouth as a one-dimensional array (vector). It should be noted that although the result of vector expression changes depending on the underlying document data and the dictionary method, it does not limit the embodiment of the present invention. 〇 2020/175371 5 卩 (: 170? 2020 /007065

There is no.

The mail data storage unit 12 stores all the mail data vectorized by the mail data analysis unit 11. Figure 3 shows an example of mail data stored in the mail data storage unit 12. The mail database stored in the mail data storage section 12 contains, for example, a mail document, vectorized data, and date and time.

The origin mail data acquisition unit 13 acquires from the mail data storage unit 1 2 the mail data that is the starting point of the thread search within the designated search period. Here, the search period and the specification of the mail data that is the starting point of the thread search are specified by the user through the user interface as shown in Fig. 4, for example.

The response candidate data generation unit 14 generates response candidate data corresponding to the origin mail data. The response candidate data generation unit 14 holds the trained model 1 4 3, and the response candidate data is generated by inputting the origin mail data to the trained model 1 4 3. This trained model 1 4 ₃ is generated by the learner 2 shown in Fig. 1. The generation of the trained model 1 4 3 in the learner 2 will be described in detail later.

The reply mail data acquisition unit 15 uses the mail data similar to the reply candidate data generated by the reply candidate data generation unit 14 as the reply mail data for replying to the origin mail data, and the mail data storage unit 1 2 To get from. When there are multiple reply mail data, multiple reply mail data are acquired. Here, the similarity between the response candidate data and the mail data stored in the mail data storage unit 12 is determined by the inner product of the vector of the response candidate data and the vector of the mail data stored in the mail data storage unit 12 Ask as. In the method of determining the degree of similarity using this vector, the method of obtaining the degree of similarity as the inner product of the vectors does not limit the embodiment of the present invention. For example, a method based on the Euclidean distance is used to compare the vectors. May be calculated.

[0019] The mail thread generation unit 16 combines the origin mail data and at least one response mail data acquired from the mail data storage unit 12 and combines them. — Generate rule thread. For example, a mail thread is created that summarizes mail exchanges as shown in Figure 5. In the example of the mail thread shown in Fig. 5, an example is shown in which the origin mail titled "Confirmation of the last day of the year" and the four reply mails that respond to it are displayed as threads.

[0020] Next, the configuration and function of the learning device 2 will be described. The learning device 2 inputs the data for e-mail learning (teaching data) as shown in Fig. 1, and passes through the learning data analysis part 21 and the data set generation part 22, and the DNN 23 learns to extract the extraction device 1 Create a trained model to be used in the response candidate data generator 14 of. The learning device 2 includes a learning data analysis unit 21, a data set generation unit 22, a DN N (deep neural network) 23, and a learned model storage unit 24. The learning data input to the learning device 2 uses a set of mail data and response mail data for the learning data as teacher data. Figure 6 shows an example of an electronic mail exchanged in the form of a question (Q u e st i o n) and a response (A n swer). Q u e st i o n and A n swer in Figure 6 are associated and extracted as pairs, and these pairs are used as teacher data. Figure 7 shows an example of a set of mail data used as teacher data and response mail data.

The learning data analysis unit 21 performs morphological analysis on the learning data and then vectorizes the learning data, similarly to the mail data analysis unit 11 included in the extraction device 1 described above. In the learning data analysis unit 21, the set of mail data and the corresponding reply mail data is converted into learning data.

The data set generation unit 22 generates a vectorized version of the input document of the teacher data and a labeled version of the response mail data as a data set to be the input of D N N 23.

[0023] DN N 23 is a deep neural network that associates vectorized incoming mail data with the label of the reply mail data. Figure 8 shows the image of learning performed in DNN 23. As shown in Fig. 8, the Q uesti 〇 n mail data subjected to morphological analysis and the response mail data corresponding to the correct answer are trained as A nswer. For example Q uestion 1 〇 2020/175371 7 卩 (: 170? 2020 /007065

The response mail data corresponding to is treated as 8 3 6 "1 and learning is performed.

.. The neural network model used in Mouth 23 is a tatami: convolutional neural network (〇 1\1) or a recursive neural network ([¾ 1\1, !_ 3 chome 1\/1). Although a neural network model such as the above can be applied, the type of the neural network model does not limit the present invention.

[0024] Next, the operation of the above-mentioned mail thread extracting device 100 will be described using a flow chart. Figure 9 shows the operation flow of the mail thread extraction device 100. In 3101, it is determined whether or not the origin mail is designated. The origin mail is specified by the user through the user interface as shown in Fig. 4, for example. The mail thread extraction device 100 obtains the origin mail data within the specified search period from the mail data storage unit 12 (3 1 0 2). Next, the mail thread extraction device 100 inputs the acquired origin mail data into the learned model 1 43 of the response candidate data generation unit 14 to generate the response candidate data (3 1 0 3) .. Then, the response mail data acquisition unit 15 acquires the response mail data similar to the response candidate data from the mail data storage unit 12 (3104). A mail thread is generated by combining the origin mail data acquired from the mail data storage unit 12 and at least one response mail data (3105). For example, the response mail data for the origin mail is output (displayed) as a mail thread as shown in Figure 5. In the example shown in Fig. 5, multiple reply mails for the mail sent at 12:00:00 on February 10, 2010, with the subject "Confirmation on the last day of the year" are sent to the mail thread. Are summarized in. In the example shown in Fig. 5, the email displayed at the top of the email thread (reply from account 8888) is selected, and the text of the email, "I am going to work normally" is displayed. , Is displayed in the right column. As described above, according to the present embodiment, at least one response mail can be collectively displayed in a thread for one origin mail.

[0025] [Second Embodiment] 〇 2020/175 371 8 卩 (: 170? 2020 /007065

As described above, one specific embodiment of the present invention has been described, but the above-described embodiment is merely an example and does not limit the present invention. For example, in the above-described first embodiment, the form in which at least one reply mail data corresponding to one origin mail is acquired and the mail thread is generated has been exemplified. However, the mail replying to the reply mail data is exemplified. The form of acquiring data may be used. Hereinafter, the mail thread extracting device according to the second embodiment of the present invention, which acquires the reply mail data corresponding to the reply mail data and generates the mail thread, will be described as follows. The different configuration will be described.

[0026] FIG. 10 shows a mail thread extraction system according to a second embodiment of the present invention.

It is a block diagram showing a schematic structure of 200. The end determination unit 17 determines whether the response mail data acquired from the mail data storage unit 12 is within the specified search period, and if the response mail data is within the search period, Send reply mail data to the origin mail data acquisition unit 13. The origin mail data acquisition unit 13 inputs the received response mail data into the response candidate data generation unit 14 and the corresponding response mail data is received in the response candidate data generation unit 14 and the response mail data acquisition unit 15 Perform the process of extracting data. This processing is performed as long as the response mail data is an email within the specified search period.

[0027] The mail thread storage unit 18 stores the origin mail data and the reply mail data. When the termination determination unit 17 determines that the response mail data is outside the search period, the termination determination unit 17 sends a termination instruction to the mail thread generation unit 16. Upon receiving the termination instruction, the mail thread generation unit 16 refers to the origin mail data and the reply mail data stored in the mail thread storage unit 18 and generates a mail thread. Note that the response mails that are judged to be out of the search period by the end judgment unit 17 are not reflected in the mail thread, and the already-stored origin mail data and reply mail data are used. To generate a mail thread.

[0028] Fig. 11 shows a flow chart of the mail thread extraction method according to the second embodiment of the present invention. 〇 2020/175 371 9 卩 (: 170? 2020 /007065

Shows mouth chart. First, wait for the search period and the starting email to be specified (3201). For example, the origin mail and search period are specified by the user using the user interface as shown in Fig. 12. The origin mail data within the specified search period is acquired from the mail data storage unit 1 2 as incoming mail data (3 2 0 2). The acquired incoming mail data is input to the trained model 1 4 3 of the response candidate data generation unit 14 to generate response candidate data (3 2 0 3). Then, the response mail data acquisition unit 15 acquires response mail data similar to the response candidate data from the mail data storage unit 12 (3 2 0 4 ). It is determined whether or not the response mail data acquired from the mail data storage unit 12 is mail data within the specified search period (3205). If the reply mail data is mail data within the specified search period, the mail thread is generated by combining the incoming mail data and the reply mail data (3206).

, Stored in the mail thread storage unit 18 (3207). Then, the reply mail data is sent to the origin mail data acquisition unit 13 as the next incoming mail data, and the processes from 3202 are repeated. On the other hand, when it is determined in the determination of 3205 that the response mail data acquired from the mail data storage unit 12 is out of the specified search period, the mail already stored in the mail thread storage unit 18 is determined. — Output the rule thread (3208). For example, in the form shown in Figure 13 the response mail data for the origin mail is collected and output as a mail thread (Table) ₀

[0029] [Third Embodiment]

In the above embodiment, the determination of the similarity between the response candidate data generated by the response candidate data generation unit 14 and the mail data stored in the mail data storage unit 12 is performed based on the response candidate data and the mail data. Although the form using the inner product of the Coutl is exemplified, a form in which a plurality of response candidate data having a high degree of similarity are extracted using a trained model trained using the dual network may be used. Hereinafter, a mail thread extraction system according to the third embodiment of the present invention will be described with respect to a configuration different from the above-mentioned embodiments. 〇 2020/175371 10 卩 (: 170? 2020 /007065

[0030] Fig. 14 shows a similarity learning model for extracting a plurality of response candidate data having a high degree of similarity. The similarity relation learning model is a learning model for extracting highly similar items from the response candidate data 8 11 3 \« 6 "1 to 8 11 3 \« 6 ". For example, a classification model Eight as candidate data

If 1 is extracted, the similarity learning model is used to

“It is possible to extract response candidate data similar to 1. By using the trained model learned by this similarity relationship learning model in the response candidate data generation unit 14, multiple response candidate data are generated. It is possible to obtain complete response candidate data in the response mail data acquisition unit 15. As a learning method for the similarity learning model, there is a teacher that uses similar response candidate data as teacher data. Although learning and unsupervised learning in which learning is performed only with response candidate data without using teacher data can be applied, the learning method does not limit the present invention.

As described above, one specific embodiment of the present invention has been described, but the above-described embodiment is merely an example and does not limit the present invention. For example, in the above embodiment, the generation of the learned model by the supervised learning is illustrated, but the learned model may be generated by the unsupervised learning. In that case, the labeling unit 24 is omitted.

Further, a part or all of the processing of each functional block of the above embodiment may be realized by a program. Then, a part or all of the processing of each functional block of each of the above-described embodiments is performed by the central processing unit (11), microprocessor, processor, etc. in the computer. Also, the programs for performing each processing are the hard disk,

Stored in a storage device such as

, Or read at 8\/1 and executed.

Further, each processing of the above-described embodiments may be realized by hardware, and includes cases where it is realized together with software (03 (operating system), middleware, or a predetermined library). ) Can be realized by 〇 2020/175371 1 1 卩 (: 170? 2020 /007065

Yes. Furthermore, the mail classification system 100 may be realized by mixed processing of software and hardware.

[0034] Further, the execution order of the processing methods in the above embodiments is not necessarily limited to the description of the above embodiments, and the execution order can be changed without departing from the gist of the invention. is there. Further, in the processing method in the above-described embodiment, some steps may be executed in parallel with other steps without departing from the scope of the invention.

A computer program that causes a computer to execute the above-described method and a computer-readable recording medium that records the program are included in the scope of the present invention. Here, the type of computer-readable recording medium is arbitrary. Further, the computer program is not limited to the one recorded in the recording medium, and may be transmitted via a telecommunication line, a wireless or wired communication line, a network typified by the internet, or the like. Good.

The specific configuration of the present invention is not limited to the above-described embodiment, and various changes and modifications can be made without departing from the gist of the invention.

The present invention can also be described as follows.

[0038] The mail thread extraction device according to the first configuration of the present invention is

A mail data storage unit for storing a plurality of email data,

A mail thread extracting unit that combines the origin mail data and the response mail data to generate a mail thread. 〇 2020/175371 12 卩 (: 170? 2020 /007065

[0039] In the first configuration, a mail document data storage unit that stores mail document data is provided, and the origin mail document data that is the origin of the mail thread is acquired from the mail document data storage unit. Then, using the learned model in which the correspondence between the mail document data and the corresponding reply mail document data is learned, the response candidate data corresponding to the starting mail document data is generated. Further, the response mail document data similar to the response candidate data is acquired from the mail document data storage unit, and the mail thread is generated based on the origin mail document data and the response mail document data. As a result, compared to the case where the related mail is extracted by the keyword search using the predetermined word and the mail thread is generated as in the conventional method, the learned model is used to deal with the mail text based on the refinement. The mail can be comprehensively extracted, and the thread extraction of the mail can be appropriately performed.

A mail thread extracting device according to a second configuration of the present invention is the mail thread extracting device according to the first configuration, wherein the response mail data acquired by the response mail data acquiring unit is the mail. It is provided with a termination determination unit that determines whether or not the response mail data is a thread target.

[0041] According to the second configuration, the terminal determination unit that determines whether or not the reply mail data is the target of the mail thread is provided, and the reply mail data is judged to be the mail of the mail thread. In this case, by inputting the response mail data into the starting point mail data acquisition unit and acquiring the mail data responding to the response mail data, the desired mail thread can be acquired without omission.

[0042] The mail thread extraction method according to the present invention is

The origin mail data, which is the origin of the mail thread, is acquired from the mail data storage section that stores multiple email data,

Using the learned model that learned the correspondence between the mail data and the corresponding reply mail data, generate the response candidate data corresponding to the origin mail data,

Store response mail data similar to the response candidate data in the mail data 〇 2020/175 371 13 卩 (: 170? 2020 /007065

Get from the department,

A mail thread is generated by combining the origin mail data and the response mail data.

[0043] According to this mail thread extraction method, the origin mail data, which is the origin of the mail thread, is acquired from the mail data storage unit that stores a plurality of e-mail data, and the mail data and the corresponding reply mail are acquired. The response candidate data corresponding to the origin mail data is generated using the learned model that has learned the correspondence with the data. Then, response mail data similar to the response candidate data is acquired from the mail data storage unit, and the start mail data and the reply mail data are combined to generate a mail thread. As a result, compared to the conventional case where related emails are extracted by keyword search using a predetermined word and an email thread is generated, it is possible to use the model that has already been trained to deal with emails that correspond to the emails based on the refinement of the email body. Can be comprehensively extracted, and the thread extraction of mail can be appropriately performed.

[0044] The program according to the present invention is

Acquiring response mail data similar to the response candidate data from the mail data storage unit,

It is a program for causing a computer to execute a process of generating a mail thread by combining the origin mail data and the response mail data.

[0045] The computer operated by this program obtains the origin mail data, which is the origin of the mail thread, from the mail database storage part that stores multiple pieces of e-mail data, and the mail data and the corresponding reply mail data. When 〇 2020/175371 14 卩 (: 170? 2020 /007065

The response candidate data corresponding to the origin mail data is generated by using the learned model in which the correspondence relationship of [1] is learned. Then, the response mail data similar to the response candidate data is acquired from the mail data storage unit, and the origin mail database and the response mail data are combined to generate a mail thread. As a result, compared to the conventional case where a related word is extracted by a keyword search using a predetermined word and a mail thread is generated, a trained model is used to search for a corresponding mail based on the e-mail text. It can be extracted comprehensively, and the thread extraction of mail can be performed appropriately.

A recording medium recording the above program is also an aspect of the present invention.

Explanation of symbols

[0047] 1 extractor, 2 learning device, 1 1 mail data analysis unit 1, 1 2 mail data storage unit, 1 3 origin mail data acquisition unit, 1 4 response candidate data generation unit, 1 5 response mail data acquisition unit, 1 6 Mail thread generation unit, 1 7 End judgment unit, 1 8 Mail thread storage unit, 2 1 Learning data analysis unit, 2 2 Data set generation unit, 2 3 0 (deep neural network), 2 4 Trained model, 100 mail thread extraction system, 200 mail thread extraction system

Claims

\¥0 2020/175 371 15 ((17 2020/007065 Claims

[Claim 1] A mail data storage unit for storing a plurality of electronic mail data,

A starting point mail data acquisition section for acquiring starting point mail data which is the starting point of the mail thread from the mail data storage section,

A response candidate data generation unit that generates response candidate data corresponding to the start point email data by using a learned model that has learned the correspondence between the email data and the corresponding response email data, and the response candidate data generation The response mail data acquisition unit that acquires the response mail data similar to the response candidate data generated from the mail data storage unit from the mail data storage unit, and the combination of the origin mail data and the reply mail data. A mail thread extraction device including a mail thread generation unit that generates a mail.

[Claim 2] The response mail data acquired by the response mail data acquisition unit is

A termination determination unit that determines whether or not the response mail data is the target of the mail thread,

The termination determination unit inputs the response mail data to the origin mail data acquisition unit when it is determined that the response mail data is response mail data to be subject to the mail thread. The mail thread extraction device described in.

[Claim 3] A mail thread extraction method executed by a computer,

Generate the response candidate data corresponding to the origin mail data by using the learned model that learned the correspondence between the mail data and the corresponding reply mail data,

The response mail data similar to the response candidate data is set to the mail data. 〇 2020/175371 16 卩 (: 170? 2020 /007065

Get from the evening storage,

A mail thread extracting method for generating a mail thread by combining the origin mail data and the response mail data.

[Claim 4] The origin mail data, which is the origin of the mail thread, is acquired from the mail data storage unit that stores a plurality of electronic mail data,

A program for causing a computer to execute a process of generating a mail thread by combining the origin mail data and the response mail data.

[Claim 5] The origin mail data, which is the origin of the mail thread, is acquired from the mail data storage unit that stores a plurality of electronic mail data,

A recording medium recording a program for causing a computer to execute a process of generating a mail thread by combining the origination mail data and the response mail data.