Mail labeling method, system, device and storage medium based on deep learning

Abstract

The present invention provides a mail labeling method, system, device and storage medium based on deep learning, the method comprising: establishing a mail data queue, collecting data messages of a first type of preset field; when the mail is a preset service for anonymously contacting a user, parsing the mail text content, grabbing the content of a second type of preset field, the second type of preset field at least including the body of the mail; pre-processing the content of the second type of preset field; based on the language of the content, inputting into a label classification model based on different languages ​​respectively, obtaining labels corresponding to the mail text content; according to the mapping relationship between the label of the mail text content and the preset label mapping table, obtaining mail processing flow information, and sending it to a preset processing mailbox. In the face of complex semantic expressions, the present invention can have higher accuracy and recall rate, facilitate timely intervention in hotel violations, and achieve timely protection of guest rights and privacy.

Description

Background Art

[0002] At present, due to the lack of effective monitoring of violations, hotels have been found to have violated regulations by cancelling full rooms, cutting off guests, and asking for guests' mobile phone numbers. This has not only resulted in guests being unable to obtain their rights and interests, privacy and security, but has also increased the company's order losses for no reason. The existing method is to use keyword extraction and T+1 review, that is, the business side reviews the suspected illegal emails from the previous day on the same day based on the results of the keyword capture, and further punishes the hotel. The keyword method is relatively simple, but the effect is poor when it comes to deep semantic expressions. Moreover, the T+1 review cannot stop the hotel's violations in a timely manner.

[0003] Therefore, the present invention provides a mail labeling method, system, device and storage medium based on deep learning. Summary of the invention

[0004] In response to the problems in the prior art, the purpose of the present invention is to provide a deep learning-based email labeling method, system, device and storage medium, which overcomes the difficulties of the prior art and can have higher accuracy and recall rate in the face of complex semantic expressions, so as to facilitate timely intervention in hotel violations and timely protect the rights and privacy of guests.

[0005] An embodiment of the present invention provides a mail labeling method based on deep learning, comprising the following steps:

[0006] S110, establishing an email data queue, collecting data messages of a first type of preset field in the email data queue, where the first type of preset field at least includes a sender's mailbox and an email subject;

[0007] S120, determining whether the email is a preset service for anonymously contacting the user, if so, executing step S130, if not, ending;

[0008] S130, parsing the email text content, and capturing the content of the second type of preset fields, where the second type of preset fields at least includes the email body;

[0009] S140, preprocessing the content of the second type of preset fields;

[0010] S150, inputting the languages ​​of the contents of the second type of preset fields into trained label classification models based on different languages ​​respectively to obtain at least one label corresponding to the email text content;

[0011] S160, obtaining email processing flow information according to the mapping relationship between the tag of the email text content and the preset tag mapping table; and

[0012] S170: Send the mail processing flow information to a preset processing mailbox.

[0013] Preferably, in step S110, the first type of preset fields includes order number, email ID, email download address, sender email address, and email subject.

[0014] Preferably, in step S120, it is determined whether the hotel is anonymously contacting the guest based on the sender's email address and the email subject.

[0015] Preferably, in the step S130, the text content of the email is parsed by HTTP request, and the second type of preset fields includes sender and receiver information, email subject, email body, and historical email conversation information.

[0016] Preferably, the step S140 includes the following steps:

[0017] S141, locating the text content of the current email according to the sender and receiver information, and filtering out the historical email conversation information;

[0018] S142, locating the email body content according to the email subject and searching for the position of a preset email opening word in the email body content;

[0019] S143, when the email language is Traditional Chinese, convert it into Simplified Chinese;

[0020] S144: When the email language is Chinese, perform Chinese word segmentation.

[0021] Preferably, in step S150, the label classification model is a pre-trained ALBERT model, and the classification labels output by the ALBERT model include fully booked, cut-out, price increase after confirmation, request for mobile phone number, and payment.

[0022] Preferably, in step S160, the preset tag mapping table includes a mapping relationship between each tag and at least one piece of mail processing flow information;

[0023] The step S170 includes processing information captured from the email corresponding to the tag, including at least the email ID, order number, email generation time, and tag.

[0024] An embodiment of the present invention further provides a deep learning-based mail label system, which is used to implement the deep learning-based mail label method described above. The deep learning-based mail label system includes:

[0025] The data queue module establishes an email data queue and collects data messages of a first type of preset fields in the email data queue, wherein the first type of preset fields at least includes a sender's mailbox and an email subject;

[0026] An anonymous judgment module determines whether the email is a preset service for anonymously contacting the user. If yes, the email parsing module is executed. If no, the process ends.

[0027] An email parsing module is used to parse the email text content and capture the content of the second type of preset fields, where the second type of preset fields at least includes the email body;

[0028] A preprocessing module, for preprocessing the content of the second type of preset fields;

[0029] A classification model module, based on the language of the content of the second type of preset field, inputs the content into a label classification model trained based on different languages ​​to obtain at least one label corresponding to the email text content;

[0030] A mapping relationship module, which obtains email processing flow information according to a mapping relationship between a label of the email text content and a preset label mapping table; and

[0031] The mail processing module sends the mail processing flow information to a preset processing mailbox.

[0032] An embodiment of the present invention further provides a mail label device based on deep learning, comprising:

[0033] processor;

[0034] a memory storing executable instructions of the processor;

[0035] Wherein, the processor is configured to perform the steps of the above-mentioned deep learning-based mail labeling method by executing the executable instructions.

[0036] An embodiment of the present invention also provides a computer-readable storage medium for storing a program, which, when executed, implements the steps of the above-mentioned deep learning-based mail labeling method.

[0037] The purpose of the present invention is to provide a mail labeling method, system, device and storage medium based on deep learning, which can build a multi-classification model based on deep learning for mail scenarios in different languages. The actual effect proves that the model has higher accuracy and recall rate in the face of complex semantic expressions. And adopt a real-time architecture to intervene in hotel violations in a timely manner, so as to timely protect the rights and interests of guests, maintain the privacy and security of guests, and reduce the company's order losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Other features, objectives and advantages of the present invention will become more apparent from a reading of the detailed description of non-limiting embodiments made with reference to the following accompanying drawings.

[0039] Figure 1 It is a flow chart of the deep learning-based mail labeling method of the present invention.

[0040] Figure 2 It is a module schematic diagram of the deep learning-based mail label system of the present invention.

[0041] Figure 3 It is a structural schematic diagram of the deep learning-based mail label device of the present invention.

[0042] Figure 4 It is a schematic diagram of the structure of a computer-readable storage medium according to an embodiment of the present invention. DETAILED DESCRIPTION

[0043] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present invention will be comprehensive and complete and fully convey the concepts of the example embodiments to those skilled in the art. The same reference numerals in the figures represent the same or similar structures, and thus their repeated description will be omitted.

[0044] Figure 1 is a flow chart of the deep learning-based email labeling method of the present invention. Figure 1 As shown, an embodiment of the present invention provides a mail labeling method based on deep learning, comprising the following steps:

[0045] S110, establishing an email data queue, collecting data messages of a first type of preset fields in the email data queue, where the first type of preset fields at least include a sender's mailbox and an email subject.

[0046] S120, determining whether the email is a preset service for anonymously contacting the user, if so, executing step S130, if not, ending.

[0047] S130, parsing the email text content, and capturing the content of the second type of preset fields, where the second type of preset fields at least include the email body.

[0048] S140: Preprocess the content of the second type of preset fields.

[0049] S150: Input the languages ​​of the contents of the second type of preset fields into trained label classification models based on different languages ​​to obtain at least one label corresponding to the email text content.

[0050] S160: Obtaining email processing flow information based on the mapping relationship between the tag of the email text content and the preset tag mapping table.

[0051] S170: Send the mail processing process information to a preset processing mailbox.

[0052] The present invention provides a real-time recognition method for multilingual email labels based on deep learning. When a hotel contacts guests through anonymous emails, the text content of the email is parsed in real time. Through a multi-classification deep learning model, hotel violations such as full room cancellation, guest cutting, and requesting guests' mobile phone numbers are captured, so as to timely protect guests' rights and interests, maintain guests' privacy and security, and reduce the company's order losses.

[0053] In a preferred embodiment, in step S110, the first type of preset fields includes order number, email ID, email download address, sender email address, and email subject.

[0054] In a preferred embodiment, in step S120, it is determined whether the hotel is anonymously contacting the guest based on the sender's email address and the email subject.

[0055] In a preferred embodiment, in step S130, the text content of the email is parsed through HTTP request, and the second type of preset fields includes sender and receiver information, email subject, email body, and historical email conversation information.

[0056] In a preferred embodiment, step S140 includes the following steps:

[0057] S141. Locate the text content of the current email based on the sender / receiver information, and filter out historical email conversation information.

[0058] S142: Locate the email body content according to the email subject and the position of the preset email opening words in the email body content.

[0059] S143. When the email language is Traditional Chinese, convert it to Simplified Chinese.

[0060] S144. When the email language is Chinese, perform Chinese word segmentation.

[0061] In a preferred embodiment, in step S150, the label classification model is a pre-trained ALBERT model, and the classification labels output by the ALBERT model include fully booked, cut-out, price increase after confirmation, request for mobile phone number, and payment.

[0062] In a preferred embodiment, in step S160, the preset label mapping table includes a mapping relationship between each label and at least one piece of mail processing flow information.

[0063] In step S170, the processing information is captured from the email corresponding to the tag, including at least the email ID, order number, email generation time, and tag.

[0064] The present invention parses the text content of emails, constructs a multi-classification model based on deep learning for email scenarios in different languages, and pushes the model discrimination results to the downstream order processing module in real time, thereby realizing effective monitoring of hotel violations in the email content, and achieving the purposes of timely protecting the guests' rights and interests, maintaining the guests' privacy and security, and reducing the company's order losses.

[0065] The present invention is mainly implemented by the following method, which mainly comprises the following steps:

[0066] The upstream sends email business data queue messages: First, it consumes the real-time data messages sent by the email subscription center to the downstream. The data is in JSON format and all required fields need to be parsed, such as email ID, email download address, sender email, email subject, etc.

[0067] Determine whether it is a business where the hotel anonymously contacts the guest: After parsing the above fields, determine whether it is a business where the hotel anonymously contacts the guest based on the sender's email address and email subject. If yes, proceed to the next step. If not, the process ends.

[0068] Parse the email text content: Then you need to parse the email text content according to the email download address through HTTP request, including sender and receiver information, email subject, email body, historical email conversation information, etc.

[0069] Text preprocessing: After parsing all the text content, the text should be preprocessed before the data enters the model. For example:

[0070] (1) Based on the sender and receiver information, locate the text content of the current email and filter out historical email conversation information.

[0071] (2) Locate the email body based on the email subject, common email opening words (Dear, My, Respectfully, Hello...), and other text content. The only data that enters the model is the email body content.

[0072] (3) If the email language is Traditional Chinese, you need to convert it to Simplified Chinese first.

[0073] (4) If the email language is Chinese, Chinese word segmentation is required first.

[0074] Multi-classification model label recognition: After completing text preprocessing, deep learning multi-classification models in different languages ​​are called according to different email languages. The ALBERT model pre-trained based on text data is used, and then fine-tuned on the email text data. Since there are multiple labels to be recognized (including: full room, cut-off, price increase after confirmation, request for mobile phone number, payment-related, etc.), two multi-classification models in Chinese and English are finally trained. In actual production use, label results are obtained through model inference.

[0075] Rule post-processing: After obtaining the model results, in order to improve the overall recognition accuracy and recall rate, a layer of rule processing is added after the model based on the special rules of the text to obtain the final recognition results.

[0076] Push results to downstream: After obtaining the label results of step 6, if the label belongs to several categories of violations, the email ID, order number, email occurrence time, model results and other fields will be pushed to the downstream in real time in json format, and then the downstream will immediately review and process it to achieve the purpose of timely protecting the guests' rights and interests, maintaining the privacy and security of the guests, and reducing the company's order losses.

[0077] The deep learning-based email labeling method of the present invention can construct a multi-classification model based on deep learning for email scenarios in different languages. The actual results show that the model has higher accuracy and recall rate in the face of complex semantic expressions. In addition, a real-time architecture is adopted to intervene in hotel violations in a timely manner, so as to timely protect the rights and interests of guests, maintain the privacy and security of guests, and reduce the company's order losses.

[0078] Figure 2 Schematic diagram of the module of the mail label system based on deep learning of the present invention. Figure 2 As shown, the embodiment of the present invention further provides a mail label system based on deep learning, which is used to implement the above-mentioned mail label method based on deep learning. The mail label system based on deep learning 5 includes

[0079] The data queue module 51 establishes an email data queue and collects data messages of a first type of preset fields in the email data queue. The first type of preset fields at least include a sender's mailbox and an email subject.

[0080] The anonymous judgment module 52 judges whether the email is a preset service for anonymously contacting the user. If so, the email parsing module is executed. If not, the process ends.

[0081] The email parsing module 53 parses the email text content and captures the content of the second type of preset fields, where the second type of preset fields at least includes the email body.

[0082] The preprocessing module 54 preprocesses the content of the second type of preset fields.

[0083] The classification model module 55 inputs the language of the content of the second type of preset field into the trained label classification models based on different languages ​​respectively to obtain at least one label corresponding to the email text content.

[0084] The mapping module 56 obtains the email processing flow information according to the mapping relationship between the tag of the email text content and the preset tag mapping table.

[0085] The mail processing module 57 sends the mail processing flow information to a preset processing mailbox.

[0086] The deep learning-based mail label system of the present invention can construct a multi-classification model based on deep learning for mail scenarios in different languages. The actual results show that the model has higher accuracy and recall rate in the face of complex semantic expressions. In addition, a real-time architecture is adopted to intervene in hotel violations in a timely manner, so as to timely protect the rights and interests of guests, maintain the privacy and security of guests, and reduce the company's order losses.

[0087] The embodiment of the present invention also provides a mail label device based on deep learning, comprising a processor, a memory, in which executable instructions of the processor are stored, wherein the processor is configured to execute the steps of the mail label method based on deep learning by executing the executable instructions.

[0088] As shown above, the deep learning-based mail label system of this embodiment of the present invention can build a multi-classification model based on deep learning for mail scenarios in different languages. The actual results prove that the model has higher accuracy and recall in the face of complex semantic expressions. And it adopts a real-time architecture to intervene in hotel violations in a timely manner, so as to timely protect the rights and interests of guests, maintain the privacy and security of guests, and reduce the company's order losses.

[0089] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as systems, methods or program products. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: complete hardware implementation, complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, which may be collectively referred to herein as "circuits", "modules" or "platforms".

[0090] Figure 3 Schematic diagram of the structure of the mail label device based on deep learning of the present invention. Figure 3 The electronic device 600 according to this embodiment of the present invention is described. Figure 3 The electronic device 600 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0091] like Figure 3 As shown, the electronic device 600 is in the form of a general computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0092] The storage unit stores program codes, which can be executed by the processing unit 610, so that the processing unit 610 executes the steps of various exemplary embodiments of the present invention described in the electronic prescription circulation processing method section of this specification. For example, the processing unit 610 can execute the following steps: Figure 1 Follow the steps shown in .

[0093] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache memory unit 6202 , and may further include a read-only memory unit (ROM) 6203 .

[0094] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0095] Bus 630 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0096] The electronic device 600 may also communicate with one or more external devices 700 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 650. Furthermore, the electronic device 600 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 660. The network adapter 660 may communicate with other modules of the electronic device 600 via a bus 630. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms, etc.

[0097] The embodiment of the present invention also provides a computer-readable storage medium for storing a program, and the steps of the deep learning-based mail label method are implemented when the program is executed. In some possible implementations, various aspects of the present invention can also be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps of various exemplary embodiments of the present invention described in the above-mentioned electronic prescription circulation processing method section of this specification.

[0098] As shown above, the deep learning-based mail label system of this embodiment of the present invention can build a multi-classification model based on deep learning for mail scenarios in different languages. The actual results prove that the model has higher accuracy and recall in the face of complex semantic expressions. And it adopts a real-time architecture to intervene in hotel violations in a timely manner, so as to timely protect the rights and interests of guests, maintain the privacy and security of guests, and reduce the company's order losses.

[0099] Figure 4 Schematic diagram of the structure of the computer-readable storage medium of the present invention. Figure 4 As shown, a program product 800 for implementing the above method according to an embodiment of the present invention is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0100] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0101] Computer readable storage media may include data signals propagated in baseband or as part of a carrier wave, wherein readable program codes are carried. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or device. The program codes contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination thereof.

[0102] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0103] In summary, the purpose of the present invention is to provide a mail labeling method, system, device and storage medium based on deep learning. The mail labeling system based on deep learning of the present invention can construct a multi-classification model based on deep learning for mail scenarios in different languages. The actual effect proves that the model has higher accuracy and recall rate in the face of complex semantic expressions. And adopt a real-time architecture to intervene in hotel violations in a timely manner, so as to timely protect the rights and interests of guests, maintain the privacy and security of guests, and reduce the company's order losses.

[0104] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.

Claims

1. A mail labeling method based on deep learning, characterized in that: The following steps are involved: S110, establishing an email data queue, collecting data messages of a first type of preset fields in the email data queue, where the first type of preset fields at least include a sender's mailbox and an email subject; S120, judging whether the email is a preset service for anonymously contacting the user according to the sender's email address and the email subject, if so, executing step S130, if not, ending; S130, parsing the email text content, and capturing the content of the second type of preset fields, where the second type of preset fields at least includes the email body; S140, preprocessing the content of the second type of preset fields; S150, inputting the languages ​​of the contents of the second type of preset fields into trained label classification models based on different languages ​​respectively to obtain at least one label corresponding to the email text content, wherein the label classification model is a pre-trained ALBERT model, and the classification labels output by the ALBERT model include fully booked, cut off, price increase after confirmation, request for mobile phone number, and payment; S160, obtaining email processing flow information according to a mapping relationship between the tag of the email text content and a preset tag mapping table; as well as S170: Send the mail processing flow information to a preset processing mailbox.

2. The deep learning-based mail labeling method according to claim 1, characterized in that: In the step S110, the first type of preset fields includes order number, email ID, email download address, sender email address, and email subject.

3. The deep learning-based mail labeling method according to claim 1, characterized in that: In the step S130, the text content of the email is parsed through HTTP request, and the second type of preset fields includes sender and receiver information, email subject, email body, and historical email conversation information.

4. The deep learning-based mail labeling method according to claim 1, characterized in that: The step S140 includes the following steps: S141, locating the text content of the current email according to the sender and receiver information, and filtering out the historical email conversation information; S142, locating the email body content according to the email subject and searching for the position of a preset email opening word in the email body content; S143, when the email language is Traditional Chinese, convert it into Simplified Chinese; S144: When the email language is Chinese, perform Chinese word segmentation.

5. The deep learning-based mail labeling method according to claim 1, characterized in that: In the step S160, the preset tag mapping table includes a mapping relationship between each tag and at least one piece of mail processing flow information.

6. A deep learning-based mail label system, used to implement the deep learning-based mail label method of claim 1, characterized in that: include: The data queue module establishes an email data queue and collects data messages of a first type of preset fields in the email data queue, wherein the first type of preset fields at least includes a sender's mailbox and an email subject; The anonymous judgment module judges whether the email is a preset service for anonymously contacting the user according to the sender's email address and the email subject. If yes, step S130 is executed; if no, the process ends; An email parsing module is used to parse the email text content and capture the content of the second type of preset fields, where the second type of preset fields at least includes the email body; A preprocessing module, for preprocessing the content of the second type of preset fields; A classification model module, based on the language of the content of the second type of preset field, inputs the content into a trained label classification model based on different languages ​​to obtain at least one label corresponding to the email text content, wherein the label classification model is a pre-trained ALBERT model, and the classification labels output by the ALBERT model include fully booked, cut off, price increase after confirmation, request for mobile phone number, and payment; A mapping relationship module, which obtains email processing flow information according to the mapping relationship between the label of the email text content and the preset label mapping table; as well as The mail processing module sends the mail processing flow information to a preset processing mailbox.

7. A mail label device based on deep learning, characterized in that: include: processor; a memory storing executable instructions of the processor; Wherein, the processor is configured to perform the steps of the deep learning-based mail labeling method according to any one of claims 1 to 5 by executing the executable instructions.

8. A computer-readable storage medium for storing a program, characterized in that: When the program is executed, the steps of the deep learning-based mail labeling method described in any one of claims 1 to 5 are implemented.

Images