CN110348009A - Email processing method, device, medium and electronic equipment based on machine learning - Google Patents

Email processing method, device, medium and electronic equipment based on machine learning Download PDF

Info

Publication number
CN110348009A
CN110348009A CN201910532896.0A CN201910532896A CN110348009A CN 110348009 A CN110348009 A CN 110348009A CN 201910532896 A CN201910532896 A CN 201910532896A CN 110348009 A CN110348009 A CN 110348009A
Authority
CN
China
Prior art keywords
mail
machine learning
unread
learning model
priority
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201910532896.0A
Other languages
Chinese (zh)
Inventor
梁锦霞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
OneConnect Smart Technology Co Ltd
Original Assignee
OneConnect Smart Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by OneConnect Smart Technology Co Ltd filed Critical OneConnect Smart Technology Co Ltd
Priority to CN201910532896.0A priority Critical patent/CN110348009A/en
Publication of CN110348009A publication Critical patent/CN110348009A/en
Priority to PCT/CN2020/087470 priority patent/WO2020253388A1/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33Querying
    • G06F16/3331Query processing
    • G06F16/334Query execution
    • G06F16/3344Query execution using natural language analysis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00Handling natural language data
    • G06F40/20Natural language analysis
    • G06F40/279Recognition of textual entities
    • G06F40/289Phrasal analysis, e.g. finite state techniques or chunking
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/045Combinations of networks
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L51/00User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail
    • H04L51/42Mailbox-related aspects, e.g. synchronisation of mailboxes

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Evolutionary Computation (AREA)
  • Databases & Information Systems (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

The embodiment of the invention provides a kind of email processing method based on machine learning, device, medium and electronic equipments, are related to technical field of data processing, this method comprises: extracting priority aspects information from the fisrt feature region of unread mail;By preset first machine learning model of priority aspects information input, the priority number of unread mail is obtained;It is greater than in the second feature region of the unread mail of preset threshold from priority number and extracts template characteristic information;By the second machine learning model of template characteristic information input, e-mail response template corresponding with unread mail is generated;E-mail response instruction is obtained, shows corresponding e-mail response template.The mail that the technical solution of the embodiment of the present invention keeps addressee's priority processing important improves the mail treatment efficiency of addressee, avoids addressee that from can not reading important email in time and miss best turnaround time.

Description

Email processing method, device, medium and electronic equipment based on machine learning
Technical field
The present invention relates to technical field of data processing, in particular to a kind of mail treatment side based on machine learning Method, device, medium and electronic equipment.
Background technique
Currently, the mail transmission/reception of mailbox and prompting etc. service in, generally can in recipient mailbox first to sender into Row grouping, according to the different type of grouping, the mail that the sender in different grouping sends is arranged to only to receive not remind or Person does not read to remind.Addressee is needed to pay special attention to and the content of preference return when existing in the mail that sender sends, but outbox People is in when only receiving the grouping that do not remind, addressee may can not in time reading mail and miss best turnaround time.
It should be noted that information is only used for reinforcing the reason to background of the invention disclosed in above-mentioned background technology part Solution, therefore may include the information not constituted to the prior art known to persons of ordinary skill in the art.
Summary of the invention
The embodiment of the present invention be designed to provide a kind of email processing method based on machine learning, device, medium and Electronic equipment, and then the mail for enabling addressee's priority processing important at least to a certain extent, improve addressee Mail treatment efficiency, avoid addressee that from can not reading important email in time and miss best turnaround time, improve addressee's E-mail response efficiency.
Other characteristics and advantages of the invention will be apparent from by the following detailed description, or partially by the present invention Practice and acquistion.
According to a first aspect of the embodiments of the present invention, a kind of email processing method based on machine learning is provided, it is described Method includes:
Priority aspects information is extracted from the fisrt feature region of unread mail;
By preset first machine learning model of the priority aspects information input, the preferential of the unread mail is obtained Series;
It is greater than in the second feature region of the unread mail of preset threshold from priority number and extracts template characteristic letter Breath;
By second machine learning model of template characteristic information input, generates mail corresponding with the unread mail and return Multiple template;
E-mail response instruction is obtained, shows the corresponding e-mail response template.
In some embodiments of the invention, aforementioned schemes are based on, the fisrt feature region includes the unread mail Mail matter topics, sender's email address, sending time and recipient mailbox address;The fisrt feature area from unread mail Priority aspects information is extracted in domain includes:
It obtains in mail matter topics, sender's email address, sending time and the recipient mailbox address of the unread mail The first text information;
The first keyword is extracted from first text information;
The priority aspects information of the unread mail is determined according to first keyword.
In some embodiments of the invention, aforementioned schemes are based on, are mentioned in the fisrt feature region from unread mail Taking out priority aspects information includes:
Obtain the characteristic area selection instruction of addressee's input;
It is determined in the unread mail according to preset characteristic area selection instruction and the fisrt feature regional location table of comparisons Fisrt feature region position;
Obtain the second text information at the corresponding fisrt feature region of the characteristic area selection instruction;
The second keyword is extracted from second text information;
The priority aspects information of the unread mail is determined according to second keyword.
In some embodiments of the invention, aforementioned schemes are based on, the fisrt feature region includes the unread mail Mail matter topics, sender's email address, sending time and recipient mailbox address;
First machine learning model obtains in the following manner:
It obtains existing with reference to the mail matter topics of mail, sender's email address, sending time and recipient mailbox address;
According to the mail matter topics with reference in mail, sender's email address, sending time and recipient mailbox address In keyword determine the existing priority aspects information with reference to mail, and generate the first training sample data;
Machine learning model is trained by the first training sample data of generation, obtains first machine Learning model.
In some embodiments of the invention, aforementioned schemes are based on, the second feature region includes the unread mail Mail matter topics, sender's email address, recipient mailbox address and message body;
Second machine learning model obtains by the following method:
It obtains existing with reference to the mail matter topics of mail, sender's email address, recipient mailbox address and message body;
According to the mail matter topics with reference in mail, sender's email address and recipient mailbox address and message body In keyword determine the existing template characteristic information with reference to mail, and generate the second training sample data;
Machine learning model is trained by the second training sample data of generation, obtains second machine Learning model.
In some embodiments of the invention, aforementioned schemes, the training sample data pair by generation are based on Machine learning model is trained, after obtaining second machine learning model further include:
Obtain modification of the addressee to e-mail response template;
Second machine learning model is trained using modified content as training sample data.
In some embodiments of the invention, aforementioned schemes are based on, the method also includes:
Multiple unread mails are arranged successively according to the descending of the priority number of the unread mail.
According to a second aspect of the embodiments of the present invention, a kind of mail treatment device based on machine learning is provided, comprising:
First extraction unit, for extracting priority aspects information from the fisrt feature region of unread mail;
Priority determining unit, for obtaining preset first machine learning model of the priority aspects information input To the priority number of the unread mail;
Second extraction unit, in the fisrt feature region from priority number greater than the unread mail of preset threshold Extract template characteristic information;
Template generation unit, for by second machine learning model of template characteristic information input, generate with it is described not Read the corresponding e-mail response template of mail;
Processing unit shows the corresponding e-mail response template for obtaining e-mail response instruction.
In some embodiments of the invention, aforementioned schemes are based on, the fisrt feature region includes the unread mail Mail matter topics, sender's email address, sending time and recipient mailbox address;
First extraction unit is configured that the mail matter topics for obtaining the unread mail, sender's email address, sends The first text information in time and recipient mailbox address;The first keyword is extracted from first text information;Root The priority aspects information of the unread mail is determined according to first keyword.
In some embodiments of the invention, aforementioned schemes are based on, first extraction unit, which is configured that, obtains addressee The characteristic area selection instruction of input;It is determined according to preset characteristic area selection instruction and the fisrt feature regional location table of comparisons The position in the fisrt feature region in the unread mail;Obtain the corresponding fisrt feature region of the characteristic area selection instruction Second text information at place;The second keyword is extracted from second text information;It is determined according to second keyword The priority aspects information of the unread mail.
In some embodiments of the invention, aforementioned schemes are based on, the fisrt feature region includes the unread mail Mail matter topics, sender's email address, sending time and recipient mailbox address;
Mail treatment device based on machine learning further include:
First acquisition unit, for obtaining the existing mail matter topics with reference to mail, sender's email address, sending time With recipient mailbox address;
First generation unit, for according to the mail matter topics with reference in mail, sender's email address, sending time The existing priority aspects information with reference to mail is determined with the keyword in recipient mailbox address, and generates the first instruction Practice sample data;
First training unit instructs machine learning model for the first training sample data by generating Practice, obtains first machine learning model.
In some embodiments of the invention, aforementioned schemes are based on, the second feature region includes the unread mail Mail matter topics, sender's email address, recipient mailbox address and message body;
The mail treatment device based on machine learning further include:
Second acquisition unit, for obtaining existing mail matter topics, sender's email address, addressee's postal with reference to mail Case address and message body;
Second generation unit, for according to the mail matter topics with reference in mail, sender's email address and addressee Keyword in email address and message body determines the existing template characteristic information with reference to mail, and generates the second instruction Practice sample data;
Second training unit instructs machine learning model for the second training sample data by generating Practice, obtains second machine learning model.
In some embodiments of the invention, aforementioned schemes are based on, the mail treatment device based on machine learning is also Include:
Third acquiring unit, the modification of the e-mail response template for obtaining addressee couple;
Third training unit is used for using modified content as training sample data to second machine learning model It is trained.
In some embodiments of the invention, aforementioned schemes are based on, the mail treatment device based on machine learning is also Include:
Sequencing unit, the descending for the priority number according to the unread mail, which is arranged successively, multiple described does not read postal Part.
According to a third aspect of the embodiments of the present invention, a kind of computer program medium is provided, computer is stored thereon with Readable instruction makes computer execute such as above-described embodiment when the computer-readable instruction is executed by the processor of computer Based on the email processing method of machine learning described in middle first aspect.
According to a fourth aspect of the embodiments of the present invention, a kind of electronic equipment is provided, comprising:
Memory is stored with computer-readable instruction;
Processor, for reading the computer-readable instruction of memory storage, to execute such as first party in above-described embodiment Based on the email processing method of machine learning described in face.
Technical solution provided in an embodiment of the present invention can include the following benefits:
In the technical solution provided by some embodiments of the present invention, by from the fisrt feature region of unread mail Priority aspects information is extracted, and by preset first machine learning model of the priority aspects information input, is not read The priority number of mail avoids addressee can not so that addressee be enable preferentially to check the higher unread mail of priority number Important mail is read in time and misses best turnaround time.In addition, by not reading postal from priority number greater than preset threshold Template characteristic information is extracted in the fisrt feature region of part, and by the second machine learning model of template characteristic information input, is given birth to At e-mail response template corresponding with unread mail, when addressee needs to reply the higher mail of priority number, lead to The instruction of acquisition e-mail response is crossed, shows corresponding e-mail response template, addressee is enable to rapidly input reply content, improves and receives The e-mail response efficiency of part people.
It should be understood that above general description and following detailed description be only it is exemplary and explanatory, not It can the limitation present invention.
Detailed description of the invention
The drawings herein are incorporated into the specification and forms part of this specification, and shows and meets implementation of the invention Example, and be used to explain the principle of the present invention together with specification.It should be evident that the accompanying drawings in the following description is only the present invention Some embodiments for those of ordinary skill in the art without creative efforts, can also basis These attached drawings obtain other attached drawings.In the accompanying drawings:
Fig. 1 diagrammatically illustrates the stream of the email processing method according to an embodiment of the invention based on machine learning Cheng Tu;
Fig. 2 diagrammatically illustrates the process according to an embodiment of the invention for obtaining priority aspects information process Figure;
Fig. 3 diagrammatically illustrates the process of acquisition priority aspects information process according to another embodiment of the invention Figure;
Fig. 4 diagrammatically illustrates the stream according to an embodiment of the invention being trained to the first machine learning model Cheng Tu;
Fig. 5 diagrammatically illustrates the stream according to an embodiment of the invention being trained to the second machine learning model Cheng Tu;
Fig. 6 diagrammatically illustrates the frame of the mail treatment device according to an embodiment of the invention based on machine learning Figure;
Fig. 7 shows the structural schematic diagram for being suitable for the computer system for the electronic equipment for being used to realize the embodiment of the present invention;
Fig. 8 shows the computer-readable of the mail treatment based on machine learning according to an embodiment of the invention and deposits Storage media figure.
Specific embodiment
Example embodiment is described more fully with reference to the drawings.However, example embodiment can be with a variety of shapes Formula is implemented, and is not understood as limited to example set forth herein;On the contrary, thesing embodiments are provided so that the present invention will more Fully and completely, and by the design of example embodiment comprehensively it is communicated to those skilled in the art.
In addition, described feature, structure or characteristic can be incorporated in one or more implementations in any suitable manner In example.In the following description, many details are provided to provide and fully understand to the embodiment of the present invention.However, It will be appreciated by persons skilled in the art that technical solution of the present invention can be practiced without one or more in specific detail, Or it can be using other methods, constituent element, device, step etc..In other cases, it is not shown in detail or describes known side Method, device, realization or operation are to avoid fuzzy each aspect of the present invention.
Block diagram shown in the drawings is only functional entity, not necessarily must be corresponding with physically separate entity. I.e., it is possible to realize these functional entitys using software form, or realized in one or more hardware modules or integrated circuit These functional entitys, or these functional entitys are realized in heterogeneous networks and/or processor device and/or microcontroller device.
Flow chart shown in the drawings is merely illustrative, it is not necessary to including all content and operation/step, It is not required to execute by described sequence.For example, some operation/steps can also decompose, and some operation/steps can close And or part merge, therefore the sequence actually executed is possible to change according to the actual situation.
Fig. 1 diagrammatically illustrates the stream of the email processing method according to an embodiment of the invention based on machine learning Cheng Tu.
Shown in referring to Fig.1, the email processing method according to an embodiment of the invention based on machine learning be can wrap Following steps S110 is included to step S150, detailed description are as follows:
In step s 110, priority aspects information is extracted from the fisrt feature region of unread mail.
In one embodiment of the invention, above-mentioned unread mail can be the unread mail in webpage mailbox, can also be with It is the unread mail in client mailbox.Priority aspects information can be in unread mail for indicating the important journey of unread mail The information of degree.
Wherein, the fisrt feature region on unread mail is the region that record has priority aspects information, the fisrt feature Region can have one or more, when the fisrt feature region of unread mail has many places, in different fisrt feature regions Priority aspects information type may not also be identical.Optionally, the fisrt feature region of unread mail may include unread mail Mail matter topics, sender's email address, sending time and recipient mailbox address etc..
In one embodiment of the invention, as shown in Fig. 2, being extracted from the fisrt feature region of unread mail preferential Grade characteristic information may include steps of S210 to step S230, and detailed description are as follows:
In step S210, mail matter topics, sender's email address, sending time and the addressee of the unread mail are obtained The first text information in people's email address.
In one embodiment of the invention, the first text information obtained in the mail matter topics of unread mail can be Particular content in mail matter topics;The first text information obtained in sender's email address and recipient mailbox address can be with For the name of sender and the name of addressee;Available sender sends the specific time of mail in the outbox time.
In step S220, the first keyword is extracted from first text information.
In one embodiment of the invention, since the acquisition region of the first text information is different, from different the The first keyword extracted in one text information is also different.Such as: it can be by the entry name in the mail matter topics of unread mail Referred to as the first keyword extraction comes out;Alternatively, when sender's remarks in the mail matter topics of unread mail have " urgent ", " weight Want ", the vocabulary such as " as early as possible reply " when, can be come out these vocabulary as the first keyword extraction;In addition, when outbox artificially compares More important people, alternatively, there are when important people in addressee, it can be by people important in addressee or sender Name come out as the first keyword extraction;Furthermore the outbox time of unread mail is more early, more addressee is needed to locate as early as possible Reason, thus using specific sending time as the first keyword extraction in the first text information that can also be obtained from the outbox time Out.
In step S230, the priority aspects information of the unread mail is determined according to first keyword.
It in one embodiment of the invention, can will be from the mail matter topics of unread mail, outbox time, sender's mailbox The first keyword extracted in address and recipient mailbox address is as priority aspects information.
Specifically, can from " project name " extracted in mail matter topics or in mail matter topics " urgent " of remarks, " important ", " replying as early as possible ";The specific time that the name of people important in sender and addressee and mail are sent As priority aspects information.
In another embodiment of the present invention, as shown in figure 3, the fisrt feature extracted region from unread mail is preferential out Grade characteristic information also may include steps of S310 to step S350, and detailed description are as follows:
In step s310, the characteristic area selection instruction of addressee's input is obtained.
In one embodiment of the invention, the fisrt feature region of unread mail is in addition to mail matter topics, outbox time, hair Outside part people email address and recipient mailbox address, addressee can also voluntarily specify the fisrt feature region of unread mail.
Wherein, different characteristic area selection instructions corresponds to the different fisrt feature regions of unread mail, and addressee passes through Inputting corresponding characteristic area selection instruction specifies some position in unread mail as fisrt feature region.
In step s 320, institute is determined according to preset characteristic area selection instruction and the fisrt feature regional location table of comparisons State the position in the fisrt feature region in unread mail.
It in one embodiment of the invention, can be by after the characteristic area selection instruction for getting addressee's input Unread mail specified by addressee is determined according to preset characteristic area selection instruction and the fisrt feature regional location table of comparisons The position in fisrt feature region.
Wherein, the region that addressee specifies can be mail matter topics, outbox time, the sender's email address of unread mail With recipient mailbox address, or a certain region of the text of unread mail, it specifically can be depending on addressee's demand.
In step S330, the second text envelope at the corresponding fisrt feature region of the characteristic area selection instruction is obtained Breath.
It in one embodiment of the invention, may be text at the corresponding fisrt feature region of characteristic area selection instruction Word, it is also possible to picture.When being text at the corresponding fisrt feature region of characteristic area selection instruction, the can be directly obtained Two text informations;When being picture at the corresponding fisrt feature region of characteristic area selection instruction, image recognition skill can be used Art is by the text conversion in picture at the second text information.
In step S340, the second keyword is extracted from second text information.
In one embodiment of the invention, since the fisrt feature region for obtaining the second text information is different, it mentions The second keyword taken out is not also identical.Such as: it can specify the text of unread mail as predeterminated position, do not read by obtaining The text of mail can be carried out when referring to addressee's name in unread mail using the name of addressee as the second keyword It extracts.
In step S350, the priority aspects information of the unread mail is determined according to second keyword.
It in one embodiment of the invention, can will be from the of the predetermined position in the fisrt feature region of unread mail The second keyword extracted in two text informations is as priority aspects information.Alternatively, second keyword can also be carried out Priority aspects information is formed after pretreatment.
It, in the step s 120, can be by preset first engineering of the priority aspects information input with continued reference to Fig. 1 Model is practised, the priority number of the unread mail is obtained.
It in one embodiment of the invention, can will be from the mail matter topics of unread mail, outbox time, sender's mailbox The fisrt feature extracted region such as address and recipient mailbox address go out priority aspects information input in advance trained machine In device learning model, the priority number of the unread mail is obtained by machine learning model.
In one embodiment of the invention, the first machine learning model is by sample training data to machine learning mould What type was trained.Wherein, machine learning model can be CNN (Convolutional Neural Network, convolution Neural network) model, or be also possible to deep neural network model etc..
As shown in figure 4, the process according to an embodiment of the invention being trained to the first machine learning model, packet Include following steps:
Step S410 is obtained existing with reference to the mail matter topics of mail, sender's email address, sending time and addressee Email address.
It in one embodiment of the invention, i.e., can be in the embodiment of the present invention with reference to mail, that is, existing mail There is mail to be analyzed, the recipient mailbox for having the theme of mail, the outbox time of existing mail, existing mail with determination Location and sender's email address.
Step S420, according to the mail matter topics with reference in mail, sender's email address, sending time and addressee Keyword in email address determines the existing priority aspects information with reference to mail, and generates the first number of training According to.
It in one embodiment of the invention, can be according to mail matter topics, sender's email address, the hair from reference mail Send the keyword that extracts in time and recipient mailbox address as first eigenvector, using the priority number of reference mail as Second feature vector generates number of training evidence.
Step S430 is trained machine learning model by the first training sample data of generation, obtains institute State the first machine learning model.
In one embodiment of the invention, first eigenvector can be inputted into the first machine learning model, obtains the The reference mail priority number of one machine learning model output, when the priority number and second feature vector with reference to mail is different When cause, the first machine learning model is adjusted, until the priority number of the reference mail of output is consistent with second feature vector.
As a result, after the completion of the training of the first machine learning model, postal can not read by the first machine learning model The priority number of part.
With continued reference to Fig. 1, in step s 130, the second spy of the unread mail of preset threshold is greater than from priority number Sign extracts template characteristic information in region.
In one embodiment of the invention, the priority of unread mail can obtained by the first machine learning model After number, the priority number of unread mail is compared with preset threshold, when the priority number of unread mail is greater than preset threshold When, then illustrate that the significance level of unread mail is higher, addressee is needed to handle as early as possible.At this point it is possible to be greater than from priority number pre- If extracting template characteristic information in the second feature region of the unread mail of threshold value.
It optionally, may include unread mail for extracting the second feature region of template characteristic information in unread mail Mail matter topics, sender's email address, recipient mailbox address and message body etc..
Wherein, template characteristic information can be project name, sender and addressee's name in mail matter topics, mail just Content in text etc..
In step S140, by second machine learning model of template characteristic information input, generates and do not read postal with described The corresponding e-mail response template of part.
It is understood that automatically generating priority number by the second machine learning model does not read postal greater than preset threshold After the e-mail response template of part, addressee can directly pass through the e-mail response template Quick-return mail after reading mail, Improve the mail treatment efficiency of addressee.
As shown in figure 5, the process according to an embodiment of the invention being trained to the second machine learning model, packet Include following steps:
Step S510, obtain it is existing with reference to the mail matter topics of mail, sender's email address, recipient mailbox address and Message body.
It in one embodiment of the invention, i.e., can be in the embodiment of the present invention with reference to mail, that is, existing mail There is mail to be analyzed, to determine the sender of the theme of existing mail, the recipient mailbox address of existing mail, existing mail The message body of email address and existing mail.
Step S520, according to the mail matter topics with reference in mail, sender's email address and recipient mailbox address, And the keyword in message body determines the existing template characteristic information with reference to mail, and generates the second training sample Data.
It in one embodiment of the invention, can be according to the mail matter topics from reference mail, sender's email address, receipts The keyword extracted in part people email address and message body is as third feature vector, by the e-mail response template of reference mail The second training sample data are generated as fourth feature vector.
Step S530 is trained machine learning model by the second training sample data of generation, obtains institute State the second machine learning model.
In one embodiment of the invention, third feature vector can be inputted into the second machine learning model, obtains the The e-mail response template of the reference mail of two machine learning models output, when the e-mail response template with reference to mail and the 4th spy When sign vector is inconsistent, the second machine learning model is adjusted, until the e-mail response template of the reference mail of output and the 4th spy It is consistent to levy vector.
As a result, after the completion of the training of the second machine learning model, postal can not read by the second machine learning model The e-mail response template of part.
With continued reference to Fig. 1, in step S150, e-mail response instruction is obtained, shows the corresponding e-mail response template.
In one embodiment of the invention, when e-mail response instruction can be the setting after addressee reads unread mail Between automatically generate, be also possible to addressee by click mail in some function button after generate.
In one embodiment of the invention, machine learning model is trained by the training sample data of generation, After obtaining the second machine learning model further include:
Modification of the addressee to e-mail response template is obtained, and using modified content as training sample data to second Machine learning model is trained.
In one embodiment of the invention, the e-mail response template that the second machine learning model generates may and not met The requirement of addressee, at this point it is possible to by addressee (namely user) to the modification content of e-mail response template, and will be in modification Hold and continue to be trained the second machine learning model as feature vector, so that the mail that the second machine learning model generates returns Multiple template meets the individual requirement of addressee.
In one embodiment of the invention, can be arranged successively according to the descending of the priority number of unread mail it is multiple not Read mail.
As a result, after addressee opens mailbox, it can make more important according to the significance level successively reading mail of mail Mail gets over advanced processing completion.
In the technical solution provided by some embodiments of the present invention, by from the fisrt feature region of unread mail Priority aspects information is extracted, and by preset first machine learning model of the priority aspects information input, is not read The priority number of mail avoids addressee can not so that addressee be enable preferentially to check the higher unread mail of priority number Important mail is read in time and misses best turnaround time.In addition, by not reading postal from priority number greater than preset threshold Template characteristic information is extracted in the second feature region of part, and by the second machine learning model of template characteristic information input, is given birth to At e-mail response template corresponding with unread mail, when addressee needs to reply the higher mail of priority number, lead to The instruction of acquisition e-mail response is crossed, shows corresponding e-mail response template, addressee is enable to rapidly input reply content, improves and receives The e-mail response efficiency of part people.
The device of the invention embodiment introduced below can be used for executing the above-mentioned mail based on machine learning of the present invention Processing method.
Fig. 6 diagrammatically illustrates the frame of the mail treatment device according to an embodiment of the invention based on machine learning Figure.
Referring to shown in Fig. 6, the mail treatment device 100 according to an embodiment of the invention based on machine learning, packet Include: the first extraction unit 110, priority determining unit 120, the second extraction unit 130, template generation unit 140 and processing are single Member 150.
Wherein, the first extraction unit 110 from the fisrt feature region of unread mail for extracting priority aspects letter Breath;Priority determining unit 120 is used to preset first machine learning model of the priority aspects information input obtaining institute State the priority number of unread mail;Second extraction unit 130 is used to be greater than from priority number the unread mail of preset threshold Second feature region in extract template characteristic information;Template generation unit 140 is used for the template characteristic information input Second machine learning model generates e-mail response template corresponding with the unread mail;Processing unit 150 is for obtaining mail Replying instruction shows the corresponding e-mail response template.
In one embodiment of the invention, the fisrt feature region includes the mail matter topics of the unread mail, hair Part people email address, sending time and recipient mailbox address;First extraction unit 110, which is configured that, obtains the unread mail Mail matter topics, sender's email address, the first text information in sending time and recipient mailbox address;From described first The first keyword is extracted in text information;The priority aspects letter of the unread mail is determined according to first keyword Breath.
In another embodiment of the present invention, the first extraction unit 110 is configured that the characteristic area for obtaining addressee's input Domain selection instruction;The unread mail is determined according to preset characteristic area selection instruction and the fisrt feature regional location table of comparisons In fisrt feature region position;Obtain the second text at the corresponding fisrt feature region of the characteristic area selection instruction Information;The second keyword is extracted from second text information;The unread mail is determined according to second keyword Priority aspects information.
In one embodiment of the invention, the fisrt feature region includes the mail matter topics of the unread mail, hair Part people email address, sending time and recipient mailbox address;The mail treatment device 100 based on machine learning also wraps It includes: first acquisition unit, it is existing with reference to the mail matter topics of mail, sender's email address, sending time and receipts for obtaining Part people's email address;First generation unit, for according to the mail matter topics with reference in mail, sender's email address, hair It send the keyword in time and recipient mailbox address to determine the existing priority aspects information with reference to mail, and generates First training sample data;First training unit, for the first training sample data by generating to machine learning mould Type is trained, and obtains first machine learning model.
In one embodiment of the invention, the second feature region includes the mail matter topics of the unread mail, hair Part people email address, recipient mailbox address and message body;The mail treatment device 100 based on machine learning also wraps It includes: second acquisition unit, for obtaining existing mail matter topics, sender's email address, recipient mailbox with reference to mail Location and message body;Second generation unit, for according to the mail matter topics with reference in mail, sender's email address and receipts Keyword in part people email address and message body determines the existing template characteristic information with reference to mail, and generates the Two training sample data;Second training unit, for the second training sample data by generating to machine learning model It is trained, obtains second machine learning model.
In one embodiment of the invention, the mail treatment device 100 based on machine learning further include: third Acquiring unit, for obtaining modification of the addressee to e-mail response template;Third training unit, for making modified content Second machine learning model is trained for training sample data.
In one embodiment of the invention, the mail treatment device 100 based on machine learning further include: sequence Unit, the descending for the priority number according to the unread mail are arranged successively multiple unread mails.
Due to example embodiments of the present invention the mail treatment device based on machine learning each functional module with it is upper The step of stating the example embodiment of the email processing method based on machine learning is corresponding, therefore in apparatus of the present invention embodiment Undisclosed details please refers to the embodiment of the processing method of the above-mentioned page resource data of the present invention.
In an exemplary embodiment of the present invention, a kind of electronic equipment that can be realized the above method is additionally provided.
Person of ordinary skill in the field it is understood that various aspects of the invention can be implemented as system, method or Program product.Therefore, various aspects of the invention can be embodied in the following forms, it may be assumed that complete hardware embodiment, complete The embodiment combined in terms of full Software Implementation (including firmware, microcode etc.) or hardware and software, can unite here Referred to as circuit, " module " or " system ".
The electronic equipment 600 of this embodiment according to the present invention is described referring to Fig. 7.The electronics that Fig. 7 is shown Equipment 600 is only an example, should not function to the embodiment of the present invention and use scope bring any restrictions.
As shown in fig. 7, electronic equipment 600 is showed in the form of universal computing device.The component of electronic equipment 600 can wrap It includes but is not limited to: at least one above-mentioned processing unit 610, at least one above-mentioned storage unit 620, the different system components of connection The bus 630 of (including storage unit 620 and processing unit 610).
Wherein, the storage unit is stored with program code, and said program code can be held by the processing unit 610 Row, so that various according to the present invention described in the execution of the processing unit 610 above-mentioned " illustrative methods " part of this specification The step of illustrative embodiments.For example, the processing unit 610 can execute step S110 as shown in fig. 1, never reading postal Priority aspects information is extracted in the fisrt feature region of part;Step S210 presets the priority aspects information input The first machine learning model, obtain the priority number of the unread mail;Step S310 is greater than preset threshold from priority number The unread mail second feature region in extract template characteristic information;Step S410, by the template characteristic information The second machine learning model is inputted, e-mail response template corresponding with the unread mail is generated;Step S510 obtains mail and returns Multiple instruction, shows the corresponding e-mail response template.
Storage unit 620 may include the readable medium of volatile memory cell form, such as Random Access Storage Unit (RAM) 6201 and/or cache memory unit 6202, it can further include read-only memory unit (ROM) 6203.
Storage unit 620 can also include program/utility with one group of (at least one) program module 6205 6204, such program module 6205 includes but is not limited to: operating system, one or more application program, other program moulds It may include the realization of network environment in block and program data, each of these examples or certain combination.
Bus 630 can be to indicate one of a few class bus structures or a variety of, including storage unit bus or storage Cell controller, peripheral bus, graphics acceleration port, processing unit use any bus structures in a variety of bus structures Local bus.
Electronic equipment 600 can also be with one or more external equipments 700 (such as keyboard, sensing equipment, bluetooth equipment Deng) communication, can also be enabled a user to one or more equipment interact with the electronic equipment 600 communicate, and/or with make Any equipment (such as the router, modulation /demodulation that the electronic equipment 600 can be communicated with one or more of the other calculating equipment Device etc.) communication.This communication can be carried out by input/output (I/O) interface 650.Also, electronic equipment 600 can be with By network adapter 660 and one or more network (such as local area network (LAN), wide area network (WAN) and/or public network, Such as internet) communication.As shown, network adapter 660 is communicated by bus 630 with other modules of electronic equipment 600. It should be understood that although not shown in the drawings, other hardware and/or software module can not used in conjunction with electronic equipment 600, including but not Be limited to: microcode, device driver, redundant processing unit, external disk drive array, RAID system, tape drive and Data backup storage system etc..
Through the above description of the embodiments, those skilled in the art is it can be readily appreciated that example described herein is implemented Mode can also be realized by software realization in such a way that software is in conjunction with necessary hardware.Therefore, according to the present invention The technical solution of embodiment can be embodied in the form of software products, which can store non-volatile at one Property storage medium (can be CD-ROM, USB flash disk, mobile hard disk etc.) in or network on, including some instructions are so that a calculating Equipment (can be personal computer, server, terminal installation or network equipment etc.) executes embodiment according to the present invention Method.
In an exemplary embodiment of the present invention, a kind of computer readable storage medium is additionally provided, energy is stored thereon with Enough realize the program product of this specification above method.In some possible embodiments, various aspects of the invention may be used also In the form of being embodied as a kind of program product comprising program code, when described program product is run on the terminal device, institute Program code is stated for executing the terminal device described in above-mentioned " illustrative methods " part of this specification according to this hair The step of bright various illustrative embodiments.
Refering to what is shown in Fig. 8, describing the program product for realizing the above method of embodiment according to the present invention 800, can using portable compact disc read only memory (CD-ROM) and including program code, and can in terminal device, Such as it is run on PC.However, program product of the invention is without being limited thereto, in this document, readable storage medium storing program for executing can be with To be any include or the tangible medium of storage program, the program can be commanded execution system, device or device use or It is in connection.
Described program product can be using any combination of one or more readable mediums.Readable medium can be readable letter Number medium or readable storage medium storing program for executing.Readable storage medium storing program for executing for example can be but be not limited to electricity, magnetic, optical, electromagnetic, infrared ray or System, device or the device of semiconductor, or any above combination.The more specific example of readable storage medium storing program for executing is (non exhaustive List) include: electrical connection with one or more conducting wires, portable disc, hard disk, random access memory (RAM), read-only Memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disc read only memory (CD-ROM), light storage device, magnetic memory device or above-mentioned any appropriate combination.
Computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal, In carry readable program code.The data-signal of this propagation can take various forms, including but not limited to electromagnetic signal, Optical signal or above-mentioned any appropriate combination.Readable signal medium can also be any readable Jie other than readable storage medium storing program for executing Matter, the readable medium can send, propagate or transmit for by instruction execution system, device or device use or and its The program of combined use.
The program code for including on readable medium can transmit with any suitable medium, including but not limited to wirelessly, have Line, optical cable, RF etc. or above-mentioned any appropriate combination.
The program for executing operation of the present invention can be write with any combination of one or more programming languages Code, described program design language include object oriented program language-Java, C++ etc., further include conventional Procedural programming language-such as " C " language or similar programming language.Program code can be fully in user It calculates and executes in equipment, partly executes on a user device, being executed as an independent software package, partially in user's calculating Upper side point is executed on a remote computing or is executed in remote computing device or server completely.It is being related to far Journey calculates in the situation of equipment, and remote computing device can pass through the network of any kind, including local area network (LAN) or wide area network (WAN), it is connected to user calculating equipment, or, it may be connected to external computing device (such as utilize ISP To be connected by internet).
In addition, above-mentioned attached drawing is only the schematic theory of processing included by method according to an exemplary embodiment of the present invention It is bright, rather than limit purpose.It can be readily appreciated that the time that above-mentioned processing shown in the drawings did not indicated or limited these processing is suitable Sequence.In addition, be also easy to understand, these processing, which can be, for example either synchronously or asynchronously to be executed in multiple modules.
Those skilled in the art after considering the specification and implementing the invention disclosed here, will readily occur to of the invention its His embodiment.This application is intended to cover any variations, uses, or adaptations of the invention, these modifications, purposes or Adaptive change follow general principle of the invention and including the undocumented common knowledge in the art of the present invention or Conventional techniques.The description and examples are only to be considered as illustrative, and true scope and spirit of the invention are by claim It points out.

Claims (10)

1. a kind of email processing method based on machine learning, which is characterized in that the described method includes:
Priority aspects information is extracted from the fisrt feature region of unread mail;
By preset first machine learning model of the priority aspects information input, the priority of the unread mail is obtained Number;
It is greater than in the second feature region of the unread mail of preset threshold from priority number and extracts template characteristic information;
By second machine learning model of template characteristic information input, e-mail response mould corresponding with the unread mail is generated Plate;
E-mail response instruction is obtained, shows the corresponding e-mail response template.
2. the method according to claim 1, wherein the fisrt feature region includes the postal of the unread mail Part theme, sender's email address, sending time and recipient mailbox address;In the fisrt feature region from unread mail Extracting priority aspects information includes:
Obtain in mail matter topics, sender's email address, sending time and the recipient mailbox address of the unread mail One text information;
The first keyword is extracted from first text information;
The priority aspects information of the unread mail is determined according to first keyword.
3. the method according to claim 1, wherein being extracted in the fisrt feature region from unread mail Priority aspects information includes:
Obtain the characteristic area selection instruction of addressee's input;
The in the unread mail is determined according to preset characteristic area selection instruction and the fisrt feature regional location table of comparisons The position of one characteristic area;
Obtain the second text information at the corresponding fisrt feature region of the characteristic area selection instruction;
The second keyword is extracted from second text information;
The priority aspects information of the unread mail is determined according to second keyword.
4. the method according to claim 1, wherein the fisrt feature region includes the postal of the unread mail Part theme, sender's email address, sending time and recipient mailbox address;
First machine learning model obtains in the following manner:
It obtains existing with reference to the mail matter topics of mail, sender's email address, sending time and recipient mailbox address;
According in the mail matter topics with reference in mail, sender's email address, sending time and recipient mailbox address Keyword determines the existing priority aspects information with reference to mail, and generates the first training sample data;
Machine learning model is trained by the first training sample data of generation, obtains first machine learning Model.
5. the method according to claim 1, wherein the second feature region includes the postal of the unread mail Part theme, sender's email address, recipient mailbox address and message body;
Second machine learning model obtains by the following method:
It obtains existing with reference to the mail matter topics of mail, sender's email address, recipient mailbox address and message body;
According in the mail matter topics with reference in mail, sender's email address and recipient mailbox address and message body Keyword determines the existing template characteristic information with reference to mail, and generates the second training sample data;
Machine learning model is trained by the second training sample data of generation, obtains second machine learning Model.
6. according to the method described in claim 5, it is characterized in that, the training sample data by generation are to machine Learning model is trained, after obtaining second machine learning model further include:
Obtain modification of the addressee to e-mail response template;
Second machine learning model is trained using modified content as training sample data.
7. the method according to claim 1, wherein the method also includes:
Multiple unread mails are arranged successively according to the descending of the priority number of the unread mail.
8. a kind of mail treatment device based on machine learning characterized by comprising
First extraction unit extracts priority aspects information from the fisrt feature region of unread mail;
Priority determining unit, for obtaining institute for preset first machine learning model of the priority aspects information input State the priority number of unread mail;
Second extraction unit, for being extracted in the second feature region from priority number greater than the unread mail of preset threshold Template characteristic information out;
Template generation unit, for generating second machine learning model of template characteristic information input and not reading postal with described The corresponding e-mail response template of part;
Processing unit shows the corresponding e-mail response template for obtaining e-mail response instruction.
9. a kind of computer program medium, is stored thereon with computer-readable instruction, when the computer-readable instruction is calculated When the processor of machine executes, computer perform claim is made to require the mail treatment based on machine learning described in any one of 1-7 Method.
10. a kind of electronic equipment characterized by comprising
Memory is stored with computer-readable instruction;
Processor, for reading the computer-readable instruction of memory storage, to execute as described in any one of claim 1-7 The email processing method based on machine learning.
CN201910532896.0A 2019-06-19 2019-06-19 Email processing method, device, medium and electronic equipment based on machine learning Pending CN110348009A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201910532896.0A CN110348009A (en) 2019-06-19 2019-06-19 Email processing method, device, medium and electronic equipment based on machine learning
PCT/CN2020/087470 WO2020253388A1 (en) 2019-06-19 2020-04-28 Machine learning-based e-mail message processing method, apparatus, medium, and electronic device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910532896.0A CN110348009A (en) 2019-06-19 2019-06-19 Email processing method, device, medium and electronic equipment based on machine learning

Publications (1)

Publication Number Publication Date
CN110348009A true CN110348009A (en) 2019-10-18

Family

ID=68182485

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910532896.0A Pending CN110348009A (en) 2019-06-19 2019-06-19 Email processing method, device, medium and electronic equipment based on machine learning

Country Status (2)

Country Link
CN (1) CN110348009A (en)
WO (1) WO2020253388A1 (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020253388A1 (en) * 2019-06-19 2020-12-24 深圳壹账通智能科技有限公司 Machine learning-based e-mail message processing method, apparatus, medium, and electronic device
US20210142150A1 (en) * 2019-11-07 2021-05-13 Fujitsu Limited Information processing device and method, and device for classifying with model
CN112818109A (en) * 2021-02-25 2021-05-18 网易(杭州)网络有限公司 Intelligent reply method, medium, device and computing equipment for mail
CN113343682A (en) * 2021-06-07 2021-09-03 中国工商银行股份有限公司 Mail processing method, mail processing device, electronic device, and storage medium
CN113592461A (en) * 2021-08-10 2021-11-02 平安普惠企业管理有限公司 Mail processing method, device and storage medium

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113408281B (en) * 2021-07-14 2024-02-09 北京天融信网络安全技术有限公司 Mailbox account anomaly detection method and device, electronic equipment and storage medium
CN113746814B (en) * 2021-08-17 2024-01-09 上海硬通网络科技有限公司 Mail processing method, mail processing device, electronic equipment and storage medium
CN114970451A (en) * 2022-04-13 2022-08-30 中国银行股份有限公司 Mail marking method, device, equipment and readable storage medium
CN116545976B (en) * 2023-07-03 2024-01-19 深圳欧税通技术有限公司 Intelligent management method, system and readable storage medium for business mail

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6816885B1 (en) * 2000-09-21 2004-11-09 International Business Machines Corporation Method and system to handle large volume of E-mail received from a plurality of senders intelligently
US7406504B2 (en) * 2003-09-18 2008-07-29 Sbc Knowledge Ventures, L.P. Intelligent email detection and auto reply email technique to emails destined to no reply email addresses
US20150200899A1 (en) * 2013-01-29 2015-07-16 Google Inc. Automatic email reminder system based on machine learning
US10911382B2 (en) * 2017-01-30 2021-02-02 Futurewei Technologies, Inc. Personalized message priority classification
CN107147779A (en) * 2017-04-21 2017-09-08 北京小米移动软件有限公司 Information cuing method and device
CN110348009A (en) * 2019-06-19 2019-10-18 深圳壹账通智能科技有限公司 Email processing method, device, medium and electronic equipment based on machine learning

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020253388A1 (en) * 2019-06-19 2020-12-24 深圳壹账通智能科技有限公司 Machine learning-based e-mail message processing method, apparatus, medium, and electronic device
US20210142150A1 (en) * 2019-11-07 2021-05-13 Fujitsu Limited Information processing device and method, and device for classifying with model
CN112818109A (en) * 2021-02-25 2021-05-18 网易(杭州)网络有限公司 Intelligent reply method, medium, device and computing equipment for mail
CN113343682A (en) * 2021-06-07 2021-09-03 中国工商银行股份有限公司 Mail processing method, mail processing device, electronic device, and storage medium
CN113592461A (en) * 2021-08-10 2021-11-02 平安普惠企业管理有限公司 Mail processing method, device and storage medium

Also Published As

Publication number Publication date
WO2020253388A1 (en) 2020-12-24

Similar Documents

Publication Publication Date Title
CN110348009A (en) Email processing method, device, medium and electronic equipment based on machine learning
CN111143535B (en) Method and apparatus for generating a dialogue model
US9865264B2 (en) Selective speech recognition for chat and digital personal assistant systems
US8326601B2 (en) Queue based mechanism to support multi-language chat rooms
CN108305626A (en) The sound control method and device of application program
US20190221208A1 (en) Method, user interface, and device for audio-based emoji input
KR102097621B1 (en) Interpreting automated assistant requests based on images and / or other sensor data
US9402576B2 (en) Electronic communication warning and modification
US11086598B2 (en) Providing a communications channel between instances of automated assistants
CN108022586A (en) Method and apparatus for controlling the page
CN108475365A (en) By management role classification and divide priority
CN111274372A (en) Method, electronic device, and computer-readable storage medium for human-computer interaction
CN108877782A (en) Audio recognition method and device
CN109791525B (en) Communication device, communication method, and storage medium
CN113674742B (en) Man-machine interaction method, device, equipment and storage medium
CN110147936A (en) Service evaluation method, apparatus based on Emotion identification, storage medium
CN107948437A (en) Put out screen and show method and apparatus
CN110516261A (en) Resume appraisal procedure, device, electronic equipment and computer storage medium
CN113785297A (en) Automatic evaluation of natural language text generated based on structured data
CN107993650A (en) Method and apparatus for generating information
CN110223694B (en) Voice processing method, system and device
CN103297611A (en) Method and system masking message on electronic device
US11656844B2 (en) Providing a communications channel between instances of automated assistants
CN105246041A (en) Prompting method and device
CN106993090A (en) Message treatment method and electronic equipment

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
CB02 Change of applicant information

Address after: 201, room 518000, building A, No. 1, front Bay Road, Qianhai Shenzhen Guangdong Shenzhen Hong Kong cooperation zone (Qianhai business secretary)

Applicant after: Shenzhen one ledger Intelligent Technology Co., Ltd.

Address before: 518000 Guangdong city of Shenzhen province Qianhai Shenzhen Hong Kong cooperation zone before Bay Road No. 1 building 201 room A

Applicant before: Shenzhen one ledger Intelligent Technology Co., Ltd.

CB02 Change of applicant information
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination