JP3742730B2 - Questionnaire mail processing system and recording medium - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a questionnaire mail processing system and a recording medium for processing an email returned by sending an email of a questionnaire.
[0002]
[Prior art]
Conventionally, an e-mail with the contents of a questionnaire as a text is sent to a member (customer), the e-mails that have been answered are temporarily accumulated, and the contents of the accumulated e-mail reply column are tabulated. At this time, the answer column is set in a standard format, and numbers are entered therein, and the numbers are mechanically recognized automatically from the column and tabulated.
[0003]
[Problems to be solved by the invention]
For this reason, when sending an email with the contents of the questionnaire to a large number of members and counting the contents of the reply fields of the large number of emails that have been replied, the numbers in the standard format are automatically recognized. However, there is a problem that emails such as complaints not described in the standard format cannot be automatically extracted and counted.
[0004]
In order to solve these problems, the present invention extracts complaint e-mails from e-mails replied at the time of questionnaires, etc., and divides the complaint contents into types to detect, aggregate, and point out complaint contents. The purpose is that.
[0005]
[Means for Solving the Problems]
Means for solving the problem will be described with reference to FIG.
In FIG. 1, a service center 1 is a center that processes a reply electronic mail by sending a questionnaire electronic mail, and is composed of a complaint extraction means 8, a complaint dictionary 9, and the like.
[0006]
The complaint extracting means 8 is for extracting an email containing a complaint with reference to the complaint dictionary 9.
The complaint dictionary 9 is for registering a complaint text.
[0007]
Next, the operation will be described.
The e-mail containing the complaint text is extracted from the e-mail sent back by the complaint extraction means 8 with reference to the complaint dictionary 9, and the extracted complaint text is stored in association with the ID of the e-mail.
[0008]
At this time, the type of complaint text is divided in advance, and the complaint text is stored for each type corresponding to the ID of the electronic mail.
Further, when displaying an electronic mail including a stored complaint text, the complaint text portion of the electronic mail is highlighted or the complaint text portion is highlighted for each type.
[0009]
Therefore, it is possible to extract a complaint e-mail from e-mails replied at the time of a questionnaire and the like, and to detect, count, and point out the complaint contents by classifying the complaint contents.
[0010]
DETAILED DESCRIPTION OF THE INVENTION
Next, embodiments and operations of the present invention will be described in detail sequentially with reference to FIGS.
[0011]
FIG. 1 shows a system configuration diagram of the present invention.
In FIG. 1, a service center 1 is a center that collects responses of e-mails sent by sending questionnaire e-mails and extracts those that contain complaints in the e-mails. The questionnaire manuscript grammar check means 2, manuscript database 3, member information database 4, questionnaire mail transmission means 5, questionnaire analysis / aggregation means 6, questionnaire reply database 7, complaint extraction means 8, and complaint dictionary 9. Is.
[0012]
The questionnaire manuscript grammar check means 2 checks the questionnaire manuscript created by the creator operating the terminal 12.
The manuscript database 3 stores checked manuscripts and registers manuscript types, titles, manuscript texts, reply deadlines, etc. in association with manuscript IDs as shown in FIG. Is.
[0013]
The member information database 4 registers member information, and registers personal information such as an e-mail address and gender in association with a member ID, as shown in FIG. is there.
[0014]
The questionnaire mail transmission means 5 creates and transmits an e-mail from the instructed member's e-mail address and the instructed manuscript.
The questionnaire analysis / aggregation means 6 aggregates data by referring to the answer column of the answered e-mail.
[0015]
The questionnaire reply database 7 is for accumulating returned electronic mails.
The complaint extracting means 8 extracts, for example, an email containing a complaint by referring to the complaint dictionary 9.
[0016]
The complaint dictionary 9 is for registering a complaint sentence in advance.
The terminal 11 is used for various operations and instructions. Here, the terminal 11 creates a questionnaire manuscript, designates a member to whom the questionnaire is sent, and instructs extraction of an e-mail including a complaint. It is something to do.
[0017]
The questionnaire member 21 is a member who has received the questionnaire e-mail, and replies to the e-mail of the questionnaire with an e-mail written in the answer column or an e-mail with a complaint. The operation of the configuration shown in FIG.
[0018]
2 and 3 are flowcharts for explaining the operation of the present invention (questionnaire mail transmission).
In FIG. 2, S1 waits for a questionnaire mail transmission instruction from the questionnaire manuscript sender.
[0019]
In S2, a questionnaire mail transmission instruction is accepted. This accepts a questionnaire mail transmission instruction with a document identifier from an operator of the terminal 11.
In step S3, a document with a specified document identifier is searched from the document database.
[0020]
In step S4, it is determined whether the search has been completed. If YES, the process proceeds to S5. In the case of NO, since there is no document with the specified document identifier in the document database 3, the survey sender is notified that the survey mail transmission instruction has failed in S6, and the process returns to S1 and is repeated.
[0021]
In S5, the document type of the specified document identifier is read from the document database.
In S6, it is determined whether the document type is a questionnaire mail. In the case of YES, since the document type of the document identifier (document ID) in FIG. 4A to be described later is found to be a questionnaire mail, the process proceeds to S8 in (2) of FIG. On the other hand, in the case of NO, since it is determined that it is not a questionnaire mail, the questionnaire sender is notified that the questionnaire mail transmission instruction has failed in S6, and the process returns to S1 and is repeated.
[0022]
In FIG. 3, a document title and document text are read from the document database in S8.
In S9, one member whose member type is a questionnaire member is searched from the member database. This searches for one member whose member type is a survey member from the member database 4 in FIG.
[0023]
In step S10, it is determined whether the search has been completed. If YES, the process proceeds to S11. In the case of NO, since there are no questionnaire members, the questionnaire mail transmission process ends and the process returns to (1).
[0024]
S11 reads the e-mail address of the member from the member database.
In step S12, an e-mail is sent to the read e-mail address with the read manuscript title as the title and manuscript text as the body. And it repeats after S9 about the next questionnaire member.
[0025]
As described above, based on the survey mail transmission instruction designating the manuscript identifier, the manuscript title and manuscript text are read from the manuscript database 3, and the survey member's e-mail address is read from the member database 4, and this is set as the e-mail destination. The scanned document title is set as the title, and the scanned document text is set as the body and transmitted. Thus, the questionnaire e-mail is transmitted to the questionnaire member.
[0026]
FIG. 4 shows an example database of the present invention.
FIG. 4A shows an example of a document database. In the manuscript database 3, the following items shown in the figure are registered.
[0027]
・ Document ID:
・ Document type: DM (direct mail), questionnaire type, etc. ・ Title:
・ Original text:
・ Return deadline: Reply deadline for questionnaire emails ・ Register the manuscript creator in the manuscript database 3 as described above, and automatically send survey e-mails to questionnaire members by specifying the manuscript ID. It becomes possible to do.
[0028]
FIG. 4B shows an example of the member information database. In the member information database 4, the following items shown in the figure are registered.
・ Member ID:
·e-mail address:
・ Personal information such as gender:
・ Member type: Survey member, DM member type, etc.
By registering member information in the member database 4 as described above, here, the survey member's e-mail address is automatically extracted to the survey member's e-mail address simply by designating the survey member (the questionnaire database 3 It is possible to automatically transmit titles and manuscript texts set as e-mail titles and texts.
[0029]
FIG. 5 shows an example of the manuscript text of the questionnaire mail of the present invention. This is an example of the manuscript text in the manuscript database 3 of FIG. 4 (a) described above. In order to conduct a questionnaire, the previous sentence (hello, △△ company △△) and the questionnaire The question (■ Q1 ..) and answer (→ A1...) Columns are provided. Here, the returned answers are automatically recognized by a predetermined character string (“■ Q1.”, “→ A1.”) At the beginning of the question and at the beginning of the answer, and the contents of the answer are analyzed and aggregated. Yes.
[0030]
6, 7 and 8 are flowcharts for explaining the operation of the present invention (questionnaire analysis / aggregation).
In FIG. 6, in S21, the questionnaire mail sender designates the questionnaire analysis. At this time, a document identifier and additional items (personal information such as gender) are designated.
[0031]
In step S22, a document with the specified document identifier is searched from the document database.
In step S23, it is determined whether the search has been completed. If YES, the process proceeds to S24. In the case of NO, since it is found that the designated document identifier does not exist in the document database, in S27, it is displayed to the questionnaire mail sender that the questionnaire analysis has failed and the process ends.
[0032]
In step S24, the document type of the specified document identifier is read from the document database.
In S25, it is determined whether the document type is a questionnaire mail. If YES, the document text of the specified document identifier is read from the document database in S26, and the process proceeds to S28 of (1) in FIG. On the other hand, in the case of NO, since it is found that the document type is not the questionnaire mail, in S27, it is displayed to the questionnaire mail sender that the questionnaire analysis has failed, and the process is terminated.
[0033]
In FIG. 7, S28 discriminates and stores the following from the read original text.
・ Total number of questions:
-Character string that is the key for each question field (first few bytes of question field):
-Character string that is the key for each answer field (first few bytes in the answer field):
・ Other:
In S29, the reply deadline of the designated document identifier is read from the document database, and the process proceeds to S30 in (2) in FIG. This reads the reply deadline date of the corresponding document ID (the deadline date for replying the questionnaire) in the document database 3 of FIG.
[0034]
In FIG. 8, in step S30, one reply mail whose reply mail identifier is the designated document identifier and whose reply time is before the reply deadline is retrieved from the questionnaire reply database.
[0035]
In step S31, it is determined whether the search has been completed. If YES, the process proceeds to S32. In the case of NO, since the search for reply e-mails targeted for all questionnaires has been completed, the process proceeds to S41.
[0036]
S32 reads the corresponding reply text from the questionnaire reply database.
In S33, a variable n = 1 and no answer flag = ON are initially set.
S34 searches for a character string to be a key of the nth question column. This searches the reply text in order from the top, and searches for the key (for example, “■ Q1.”) In the nth (n = 1, 2,...) Question column.
[0037]
In S35, it is determined whether the search has been made. In the case of YES, since the nth question column can be found from the reply mail, the process proceeds to S36. If NO, the process proceeds to S38.
In S36, a character string to be a key of the nth answer field is searched.
[0038]
In S37, it is determined whether the search has been made. In the case of YES, the process proceeds to S43 of (4) in FIG. If NO, the process proceeds to S38.
In S38, it is determined whether n is 1. In the case of YES, the number that could not be analyzed in S39 is counted up, and the process proceeds to S49 of (5) in FIG. On the other hand, in the case of NO, the number that can be analyzed halfway in S40 is counted up, and the process proceeds to S49 of (5) in FIG.
[0039]
In FIG. 9, in S43, it is determined whether the member describes the answer information in the answer column. If YES, the answer information is stored in S44, the no answer flag is turned OFF, and the process proceeds to S45. On the other hand, if NO, the process proceeds to S45.
[0040]
In S45, it is determined whether n is the total number of question fields. If YES, it is determined in S46 whether the no answer flag is ON. If YES, the number of all unanswered questions is counted up in S47, and the process proceeds to S49. If NO, the number of all questions that can be analyzed in S48. Is counted up, and the process proceeds to S49. On the other hand, if NO in S45, n is incremented in S53, and the process returns to S34 in FIG.
[0041]
In S49, it is determined whether additional information is designated. If YES, the process proceeds to S50. If no, the process proceeds to S52.
In S50, the member number of the responder is read from the questionnaire reply database.
[0042]
S51 reads the personal information (gender, etc.) of the member number from the member information database.
In S52, the following information for one mail is added to the result list. Then, the process returns to S30 of FIG.
[0043]
・ Reply information in all response fields (indication of that in the fields where responses could not be made):
-Response results (all questions can be analyzed, etc.):
・ Personal information (when additional items are specified):
・ Other:
By the above, search the key words in the question column in the questionnaire reply mail to find the question, search the key characters in the answer column to find the answer, and store the answer information in the found answer column Repeatedly, the response information in all response fields and the number of responses / no responses are automatically counted, and the results (response information in all response columns, total responses / total no responses, personal information, etc.) are automatically calculated. It becomes possible to output.
[0044]
FIG. 10 shows a result list example at the time of questionnaire analysis of the present invention. This is an example of a result list obtained by analyzing and counting reply mails according to the flowcharts of FIGS. 7 to 9 described above, and is a list of the following items shown in the figure.
[0045]
・ Member ID:
・ Additional information (personal information): Gender, age, etc. ・ Q1 answer:
・ Answer of Q2:
...
・ Qn's answer:
・ Analysis results: 〇 (all questions were analyzed), △ (analyzed halfway), × (no single question was analyzed), □ (all questions not answered)
・ Other:
In the figure, Q1 answer, Q2 answer... In * 1 indicates that analysis was not possible, and * 2 indicates that analysis was possible but nothing was written in the answer column.
[0046]
FIG. 11 shows another system configuration diagram of the present invention.
In FIG. 11, a program 31 that operates periodically is a program that operates and processes periodically, and includes a questionnaire response abnormality detection means 32, a complaint wording detection means 33, and a comprehensive determination means 34 as shown in the figure. It is what is done.
[0047]
The questionnaire reply abnormality detecting means 32 is for detecting an abnormal email reply to the questionnaire with reference to the manuscript database 3 with respect to the electronic mail retrieved from the questionnaire reply database 7 and returned.
[0048]
The complaint wording detection means 33 detects a complaint text of an email returned based on the questionnaire reply database 7, the manuscript database 13, and the complaint dictionary 35.
[0049]
The comprehensive determination means 34 performs a comprehensive determination.
The complaint dictionary 35 is a dictionary in which a complaint character string is registered in advance.
The complaint candidate list 36 is a list of complaint e-mail candidates in the replied e-mail.
[0050]
Next, FIG. 12 and FIG. 13 show an operation explanation flowchart (complaint candidate list) of the present invention.
In FIG. 12, S61 retrieves one document identifier whose document type is questionnaire mail from the document database.
[0051]
In step S62, it is determined whether the search has been completed. If YES, the process proceeds to S63. In the case of NO, since the search was not possible, the process is terminated.
In S63, the document text with the document identifier is read from the document database.
[0052]
In S64, the following is discriminated from the read original text and stored.
・ Total number of questions:
-Character string that is the key for each question field (first few bytes of question field):
-Character string that is the key for each answer field (first few bytes in the answer field):
・ Other:
In S65, one reply identifier whose reply mail manuscript identifier is the above identifier and whose complaint processing flag is unprocessed is searched from the questionnaire reply database.
[0053]
In S66, it is determined whether the search has been completed. In the case of YES, the process proceeds to S67 of FIG. If NO, the process returns to S61 and repeats for the processing of the next document.
In FIG. 13, S67 reads the corresponding reply text from the questionnaire reply database.
[0054]
In S68, variables n = 1 and no answer flag = ON are initialized.
In S69, a character string to be a key of the nth question column is searched.
In step S70, it is determined whether the search has been made. If YES, the process proceeds to S71. In the case of NO, the pair of the document identifier and the reply identifier is written in the complaint candidate list in S77, and the process returns to S65 in FIG.
[0055]
In S71, a character string to be a key of the nth answer field is searched.
S7 3 determines whether members describe the answer information in response column. If YES, the free response flag to OFF in S7 4, the process proceeds to S7 5. On the other hand, if NO, the process proceeds to S7 5.
[0056]
S7 5, n is determined whether all questions number. In the case of YES, because you are done for all the questions, it proceeds to S7 6. In the case of NO, counts up the n in S7 7, repeat the S69 and later for the next n-th.
[0057]
S7 6 is No Answer flag determines whether ON. If YES, S7 8 writes the pairs of document identifiers and reply identifier complaints candidate list, repeated returns to S65 of ○ 2 of Figure 12. On the other hand, if NO, the process returns to S65 of FIG.
[0058]
Search the question field and answer field for the reply e-mail as described above. ・ When the question field cannot be found ・ When the answer field cannot be found ・ When the answer information is not described in the answer field, etc. It is possible to automatically register a pair of a document identifier and a reply identifier.
[0059]
14 and 15 are flowcharts for explaining the operation of the present invention (complaint extraction).
In FIG. 14, S81 searches the document database for one document identifier whose document type is questionnaire mail.
[0060]
In step S82, it is determined whether the search has been completed. If YES, the process proceeds to S83. If NO, the process ends.
In step S83, the document text having the document identifier is read from the document text.
[0061]
In step S84, it is determined whether or not all the complaint expressions contained in the complaint dictionary are used in the manuscript text, and a temporary complaint dictionary is created from what is excluded from the used complaint expressions.
[0062]
In S85, one reply identifier whose reply mail manuscript identifier is the above-mentioned manuscript identifier and whose complaint processing flag is unprocessed is searched from the questionnaire reply database.
In step S86, it is determined whether the search has been completed. In the case of YES, the process proceeds to S87 of FIG. If NO, return to S81 and repeat.
[0063]
In FIG. 15, S87 reads the corresponding reply text from the questionnaire reply database.
S88 determines whether at least one complaint wording in the temporary complaint dictionary is used in the reply text.
[0064]
S89 determines whether it was used. In the case of YES, the pair of the document identifier and the reply list is written in the complaint candidate list in S90 (if the same information has already been written, it is not written), and the process returns to S85 in FIG. Repeat about. On the other hand, in the case of NO in S89, the process returns to S85 in FIG.
[0065]
As described above, it is possible to extract an e-mail from the e-mail that has a character string for complaining that is not included in the manuscript text and write it in the complaint candidate list.
FIG. 16 shows an example of a complaint dictionary of the present invention. The complaint dictionary 9 registers the following items shown in the figure.
[0066]
-Complaint ID:
-Complaint type: Word, ending, etc.-Complaint terms:
・ Other:
FIG. 17 shows a complaint candidate list example (before comprehensive judgment) of the present invention. This complaint candidate list is the one before the comprehensive determination, and the following items shown are registered (written and registered according to the flowcharts of FIGS. 12 to 15).
[0067]
・ Document identifier:
・ Reply list:
・ Questionnaire reply abnormality analysis results:
-Complaining word ending-Complaining wording-Other:
Here, * 1 is a part detected by the questionnaire reply abnormality detecting means * 2 is a part detected by the complaint wording detecting means * 3 is a combination of the same manuscript identifier / reply list after the comprehensive judging means is operated. Therefore, either one is deleted.
[0068]
FIG. 18 shows an operation explanation flowchart (overall determination) of the present invention.
In FIG. 18, S101 reads one complaint candidate list. This is read from the complaint candidate list of FIG.
[0069]
In step S102, other combinations of the same document identifier / reply list are searched. This searches for another identical document identifier / reply list pair of one complaint information read from the complaint candidate list of FIG. For example, it is searched whether there are two identical items * 3 in FIG.
[0070]
In step S103, it is determined whether the search has been completed. In the case of YES, any one is deleted in S104 (any one of * 3 in FIG. 7 is deleted), and the process proceeds to S105. If NO, the process proceeds to S105.
[0071]
In step S105, it is determined whether all the complaint candidate lists have been read. If YES, the process ends. If no, return to S101 and repeat for the next complaint candidate.
As described above, the duplicate combination of the document identifier / reply list is deleted from the complaint candidate list of FIG. 17, and a comprehensive determination (total determination such as the total number of complaint candidates for the questionnaire, the total number of complaint candidates for each document identifier) is performed. It becomes possible. In addition, based on the list of complaint candidates, the complaint part of the e-mail sent back to the center's questionnaire creator is highlighted and presented for each type, or the total result for each type of complaint is displayed. It becomes possible.
[0072]
【The invention's effect】
As described above, according to the present invention, a complaint e-mail is extracted from an e-mail replied at the time of a questionnaire and the like, and the complaint contents are classified into types and detected and the complaint contents are classified into highlights. Therefore, it is possible to automatically extract and highlight a complaint from an e-mail replied at the time of the questionnaire and to indicate it or to sum up the contents of the complaint.
[Brief description of the drawings]
FIG. 1 is a system configuration diagram of the present invention.
FIG. 2 is a flowchart for explaining the operation of the present invention (questionnaire mail transmission, part 1).
FIG. 3 is a flowchart for explaining the operation of the present invention (questionnaire mail transmission, part 2).
FIG. 4 is a database example of the present invention.
FIG. 5 is an example of a manuscript text of a questionnaire mail according to the present invention.
FIG. 6 is a flowchart for explaining the operation of the present invention (questionnaire analysis / aggregation, part 1).
FIG. 7 is a flowchart for explaining the operation of the present invention (questionnaire analysis / aggregation, part 2).
FIG. 8 is a flowchart for explaining the operation of the present invention (questionnaire analysis / aggregation, part 3).
FIG. 9 is a flowchart for explaining the operation of the present invention (questionnaire analysis / aggregation, part 4).
FIG. 10 is a result list example at the time of questionnaire analysis of the present invention.
FIG. 11 is another system configuration diagram of the present invention.
FIG. 12 is a flowchart for explaining the operation of the present invention (complaint candidate list, part 1).
FIG. 13 is a flowchart for explaining the operation of the present invention (complaint candidate list, part 2).
FIG. 14 is a flowchart for explaining the operation of the present invention (complaint extraction, part 1).
FIG. 15 is a flowchart for explaining the operation of the present invention (complaint extraction, part 2).
FIG. 16 is an example of a complaint dictionary according to the present invention.
FIG. 17 is an example of a complaint candidate list (before comprehensive judgment) according to the present invention.
FIG. 18 is a flowchart for explaining the operation of the present invention (overall determination).
[Explanation of symbols]
1: Service Center 2: Questionnaire Manuscript Grammar Checking Unit 3: Manuscript Database 4: Member Information Database 5: Questionnaire Mail Transmission Unit 6: Questionnaire Analysis / Aggregation Unit 7: Questionnaire Reply Database 8: Complaint Extraction Unit 9, 35: Complaint Dictionary 11 : Terminal 21: questionnaire member 32: questionnaire reply abnormality detection means 33: complaint wording detection means 34: comprehensive judgment means 36: complaint candidate list

Claims (6)

In the questionnaire mail processing system that processes the reply mail for the email of the sent questionnaire,
Means for receiving and storing reply emails;
First, a temporary complaint dictionary is created and stored for each manuscript identifier of the questionnaire, excluding the complaint wording used in the transmitted questionnaire email, from among the complaint dictionaries in which the complaint words are registered in advance. Storage means,
Referring to the temporary complaints dictionary, among the reply mail that the storage, means for extracting an e-mail including complaints sentence from the same document identifier a is in the reply mail and document ID of the questionnaire,
A questionnaire mail processing system, comprising: a second storage unit that stores the extracted complaint text in association with the ID of the reply mail that includes the complaint text.   2. A questionnaire mail processing system according to claim 1, wherein the complaint text is divided into types, and the second storage means stores the complaint text in association with the ID of the reply mail for each type. .   When displaying the reply mail including the complaint text stored in the second storage means, the complaint text section of the reply mail is highlighted or the complaint text section is highlighted for each type. The questionnaire mail processing system according to claim 1 or 2, wherein: A program for causing a computer to process a reply e-mail to a sent questionnaire e-mail,
Computer
Means for receiving and storing reply emails;
First, a temporary complaint dictionary is created and stored for each manuscript identifier of the questionnaire, excluding the complaint wording used in the transmitted questionnaire email, from among the complaint dictionaries in which the complaint words are registered in advance. Storage means,
Referring to the temporary complaints dictionary, among the reply mail that the storage, means for extracting an e-mail including complaints sentence from the same document identifier a is in the reply mail and document ID of the questionnaire,
A computer-readable recording medium storing a program that functions as a second storage unit that stores the extracted complaint text in association with the ID of the reply mail that contained the complaint text. In the above questionnaire email processing system,
The first storage means discriminates and stores a character string to be a key of each answer column for each manuscript identifier of the questionnaire from the transmitted questionnaire email.
The means for extracting searches the reply field based on the character string for each mail from reply mails having the same document identifier as that of the questionnaire in the accumulated reply mails, Extract the answer information described in the answer column that was searched,
2. The questionnaire mail processing system according to claim 1, wherein the second storage means stores the extracted answer information in association with an ID of a reply mail from which the answer information has been extracted. The first storage means further determines and stores the total number of questions from the email of the transmitted questionnaire,
The means for extracting further counts the number of response fields that can be searched based on the character string for each mail from the accumulated reply mails.
The second storage means further stores information indicating whether or not the answer information of the reply mail has been analyzed based on the counted number of answer fields and the total number. The questionnaire mail processing system according to claim 5.

Images