CN108922634A - The problem of based on online interrogation, replies processing method, device and computer equipment - Google Patents
The problem of based on online interrogation, replies processing method, device and computer equipment Download PDFInfo
- Publication number
- CN108922634A CN108922634A CN201810606392.4A CN201810606392A CN108922634A CN 108922634 A CN108922634 A CN 108922634A CN 201810606392 A CN201810606392 A CN 201810606392A CN 108922634 A CN108922634 A CN 108922634A
- Authority
- CN
- China
- Prior art keywords
- user
- vector
- doctor
- text
- question
- 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
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 12
- 239000013598 vector Substances 0.000 claims abstract description 129
- 238000004891 communication Methods 0.000 claims abstract description 58
- 238000000034 method Methods 0.000 claims abstract description 28
- 238000012545 processing Methods 0.000 claims description 41
- 239000011159 matrix material Substances 0.000 claims description 21
- 230000015654 memory Effects 0.000 claims description 20
- 238000012512 characterization method Methods 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 17
- 210000005036 nerve Anatomy 0.000 claims description 12
- 238000012544 monitoring process Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 238000003745 diagnosis Methods 0.000 description 3
- 208000024891 symptom Diseases 0.000 description 3
- 206010019233 Headaches Diseases 0.000 description 2
- 231100000869 headache Toxicity 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 206010029421 Nipple pain Diseases 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000002560 therapeutic procedure Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H80/00—ICT specially adapted for facilitating communication between medical practitioners or patients, e.g. for collaborative diagnosis, therapy or health monitoring
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Medical Informatics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Epidemiology (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Public Health (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Pathology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The problem of being based on online interrogation this application involves one kind replies processing method, device and computer equipment.The method includes：Receive the interrogation request of user terminal uploads；The interrogation request carries user identifier；It is that the user identifier distributes corresponding doctor according to interrogation request, establishes the communication connection between the user terminal and doctor terminal；Monitor the communication information between the user terminal and the doctor terminal；When listening to user's enquirement in the communication information, calculates the user and put question to corresponding text vector；The corresponding text vector of sample problem is obtained, the corresponding text vector of user enquirement text vector corresponding with the sample problem is matched, is filtered out in the sample problem and the user puts question to matched multiple recommendation problems and recommends answer；By multiple recommendation problems and answer is recommended to be sent to corresponding doctor terminal.The time-consuming for replying user and puing question to can effectively be shortened using this method, improve and reply accuracy.
Description
Technical field
This application involves field of computer technology, in particular to the problem of one kind replies processing side based on online interrogation
Method, device, computer equipment and storage medium.
Background technique
It can be convenient user by online interrogation platform to seek advice from regard to disease problems are online Xiang doctor.User propose problem it
Afterwards, it is desirable to access doctor and timely reply.And doctor needs after seeing the enquirement of user by carefully thinking deeply again
It is manually entered answer content.If the problem of user is complex, the time that doctor replies may be longer.Since user proposes
The problem of there is a large amount of similar and duplicate, a large amount of problems proposed in common mode using user and doctor
FAQs library is established in answer, and when doctor is in the enquirement for receiving user, suitable problem can be retrieved in FAQs library,
The corresponding content that replies is sent to user, replies the time-consuming that user puts question to shorten doctor.
But this mode is retrieved the problem of obtaining and may and be mismatched practical the problem of proposing with user.For example, retrieval
The problem related to " headache ", what may be retrieved is " what if is nipple pain ".In this case, it may be necessary to repeated retrieval or
Person doctor is manually entered answer content.Therefore, how effectively to shorten reply user put question to time-consuming, improve reply accuracy at
For a technical problem for needing to solve at present.
Summary of the invention
Based on this, it is necessary in view of the above technical problems, provide a kind of can effectively shorten and reply the consumption that user puts question to
When, improve reply accuracy based on online interrogation the problem of reply processing method, device, computer equipment and storage medium.
One kind replying processing method based on the problem of online interrogation, the method includes：
Receive the interrogation request of user terminal uploads；The interrogation request carries user identifier；
It is that the user identifier distributes corresponding doctor according to interrogation request, establishes the user terminal and doctor is whole
Communication connection between end；
Monitor the communication information between the user terminal and the doctor terminal；
When listening to user's enquirement in the communication information, calculates the user and put question to corresponding text vector；
The corresponding text vector of sample problem is obtained, puts question to the user to corresponding text vector and the sample problem
Corresponding text vector is matched, and is filtered out in the sample problem and is putd question to matched multiple recommendation problems with the user
With recommendation answer；
By multiple recommendation problems and answer is recommended to be sent to corresponding doctor terminal.
It is described in one of the embodiments, to distribute corresponding doctor's packet according to interrogation request for the user identifier
It includes：
A variety of user informations are obtained according to the user identifier；
Vectorization processing is carried out to the user information, obtains multi-C vector matrix；
The multi-C vector matrix is based on by first nerves network model and carries out prediction operation, obtains marking with the user
Know corresponding department；
Corresponding doctor is distributed to the user identifier according to the department.
The corresponding text vector of user's enquirement that calculates includes in one of the embodiments,：
The user is putd question to and is pre-processed, pretreated text is obtained；
Nervus opticus network model is called, the nervus opticus network model includes LSTM and full articulamentum；
Operation is carried out to pretreated text by the LSTM, obtains characterization vector；
The characterization vector is converted by the full articulamentum, obtain text corresponding with user enquirement to
Amount.
The LSTM includes forgeing door, input gate and out gate in one of the embodiments, described to pass through the LSTM
Carrying out operation to pretreated text includes：
When getting the pretreated text at current time, by forgeing door to the pretreated text of last moment
This carries out forgetting processing；
When the forgetting door carries out forgeing processing, current time is inputted by the input gate pretreated text
Originally it is updated；
Operation is carried out by the text that out gate obtains obtained text after forgetting processing and update, is obtained and user
Put question to corresponding characterization vector.
It is described in one of the embodiments, to put question to corresponding text vector corresponding with the sample problem user
Text vector carry out matching include：
Calculate the user put question to it is similar between corresponding text vector text vector corresponding with multiple sample problems
Degree；
When the similarity reaches threshold value, the corresponding sample problem of the similarity is labeled as problem to be recommended；
Screened in multiple problems to be recommended, by the problem flag to be recommended of the preset quantity filtered out be with it is described
User puts question to corresponding recommendation problem.
One kind replying processing unit based on the problem of online interrogation, and described device includes：
Communication module, the interrogation for receiving user terminal uploads are requested；The interrogation request carries user identifier；
Doctor's distribution module, for being that the user identifier distributes corresponding doctor according to interrogation request；
The communication module is also used to establish the communication connection between the user terminal and doctor terminal；
Module is monitored, for monitoring the communication information between the user terminal and the doctor terminal；
Vectorization processing module when puing question to for listening to user in the communication information, calculates the user and puts question to
Corresponding text vector；
Matching module puts question to the user to corresponding text vector for obtaining the corresponding text vector of sample problem
Text vector corresponding with the sample problem is matched, and is filtered out in the sample problem and is matched with user enquirement
Multiple recommendation problems and recommend answer；
The communication module is also used to multiple recommendation problems and answer is recommended to be sent to corresponding doctor terminal.
User identifier is carried in the interrogation request in one of the embodiments,；Described device further includes：
Doctor's distribution module, for obtaining a variety of user informations according to the user identifier；The user information is carried out
Vectorization processing, obtains multi-C vector matrix；The multi-C vector matrix is based on by first nerves network model to be predicted
Operation obtains department corresponding with the user identifier；Corresponding doctor is distributed to the user identifier according to the department.
The vectorization processing module is also used to put question to the user and pre-process in one of the embodiments,
Obtain pretreated text；Nervus opticus network model is called, the nervus opticus network model includes LSTM and connects entirely
Connect layer；Operation is carried out to pretreated text by the LSTM, obtains characterization vector；By the full articulamentum to described
Characterization vector is converted, and text vector corresponding with user enquirement is obtained.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
Device realizes the step in above-mentioned each embodiment of the method when executing the computer program.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
The step in above-mentioned each embodiment of the method is realized when row.
Above-mentioned the problem of being based on online interrogation, replies processing method, device, computer equipment and storage medium, when user needs
When wanting online interrogation, can be requested according to interrogation in the user identifier that carries be that user distributes corresponding doctor, establish user's end
Communication connection between end and doctor terminal, and the communication information between user terminal and doctor terminal is monitored.?
When listening to user's enquirement in the communication information, calculates the user and put question to corresponding text vector, the user is putd question to
Corresponding text vector text vector corresponding with the sample problem is matched, and is filtered out in sample problem and the use
Family puts question to matched multiple recommendation problems and recommends answer, and it is whole that multiple recommendation problems and recommendation answer are sent to corresponding doctor
End.During making online interrogation, is putd question to without doctor for user and think deeply and be manually entered answer, it can
To select to put question to immediate recommendation problem with user directly from recommendation problem, answer is recommended to be sent to user's end by corresponding
End.To effectively shorten the time-consuming for replying user and puing question to.Since recommendation problem is to match to obtain by carrying out text vector
, recommendation problem is putd question to user and is close, and is putd question to using recommending answer to reply user, is effectively increased the accuracy of answer.
Detailed description of the invention
The problem of Fig. 1 is in one embodiment based on online interrogation replies the application scenario diagram of processing method；
The problem of Fig. 2 is in one embodiment based on online interrogation replies the flow diagram of processing method；
Fig. 3 be in one embodiment according to interrogation request be user identifier distribute corresponding doctor's step process illustrate
Figure；
Fig. 4 is in one embodiment by the corresponding text vector of user enquirement text corresponding with the sample problem
The flow diagram of vector progress matching step；
The problem of Fig. 5 is in one embodiment based on online interrogation replies the structural block diagram of processing unit；
Fig. 6 is the internal structure chart of computer equipment in one embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
The problem of being based on online interrogation provided by the present application, replies processing method, can be applied to application as shown in Figure 1
In environment.Wherein, user terminal 102 is communicated by network with server 104.Doctor terminal 106 passes through network and service
Device 104 is communicated wherein, and user terminal 102 and doctor terminal 106 can be, but not limited to be various personal computers, notes
This computer, smart phone, tablet computer and portable wearable device, server 104 can be with independent servers either
The server cluster of multiple servers composition is realized.When user wishes on-line consulting doctor, user terminal 102 can be passed through
Interrogation request is uploaded to server 104.Server 104 is requested according to interrogation, distributes corresponding doctor for user, establishes user's end
Communication connection between end 102 and doctor terminal 106.Server 104 is to the communication information between user terminal and doctor terminal
It is monitored.In snoop procedure, if recognized in the communication information, there are user's enquirements, calculate user and put question to corresponding text
This vector.Problem base is set up on server 104, includes a variety of sample problems in problem base.Server 104 obtains sample
User is putd question to corresponding text vector text vector corresponding with sample problem to match by the corresponding text vector of problem,
It is filtered out in sample problem and puts question to matched multiple recommendation problems with user and recommend answer.Server 104 is by multiple recommendations
Problem and recommendation answer are sent to corresponding doctor terminal 106.Doctor is selected in multiple recommendation problems by doctor terminal 106
Immediate recommendation problem is putd question to user and recommends answer, and the recommendation answer selected is sent to the user terminal 102.
In one embodiment, processing method is replied based on the problem of online interrogation as shown in Fig. 2, providing one kind, with
This method is applied to be illustrated for the server in Fig. 1, includes the following steps：
Step 202, the interrogation request of user terminal uploads is received；Interrogation request carries user identifier.
Step 204, be that user identifier distributes corresponding doctor according to interrogation request, establish user terminal and doctor terminal it
Between communication connection.
The application program of online interrogation is mounted on user terminal.Before logging in the application program, user needs to carry out
Registration.The essential informations such as user identifier, gender are contained in registration information.When user wishes on-line consulting doctor, Ke Yidian
The interrogation button in application program is hit, interrogation request is generated, interrogation request is sent to server by application program.Server connects
The interrogation request for receiving user terminal uploads, calls first nerves network model, is user's distribution by first nerves network model
Department.Server is that user distributes corresponding doctor according to department.Server is established logical between user terminal and doctor terminal
Letter connection.
Step 206, the communication information between monitoring users terminal and doctor terminal.
It is established after communication connection between user terminal and doctor terminal, user inputs main suit by user terminal and believes
Breath.It include content relevant to user's symptom in main suit's information.Main suit's information is sent to doctor terminal by user terminal.Doctor
It can be putd question to by doctor terminal to user, to understand the details of user.User can also pass through user terminal
It sends and puts question to doctor terminal.When carrying out real time communication between user terminal and doctor terminal, server can pass through monitoring
Script monitors the communication information between user terminal and doctor terminal, puts question to automatically identify user.
Step 208, it when recognizing user's enquirement in the communication information, calculates user and puts question to corresponding text vector.
When being communicated between user terminal and doctor terminal, that use can be carried in the communication information of user terminal
Family mark or terminal iidentification.Doctor identification or doctor terminal mark can be carried in the communication information from doctor terminal
Know.When server monitors the communication information between user terminal and doctor terminal, can identify in the communication information whether
There are user's enquirements.Specifically, server can carry out the communication information listened to according to preset characters or preset characters string
Matching, if identifying in the communication information listened to there are preset characters or preset characters string, and the communication information is used by oneself
Family terminal, then server puts question to this communication information labeled as user.For example, preset characters be "？", preset characters string is
" how ", " how " etc..The wherein communication information listened to is " recently always headache this how to treat ", and server exists
Had identified in the communication information preset characters string " how ", and carry user identifier in the communication information, indicate the information
From client terminal, then server can put question to this communication information labeled as user.
Server calls nervus opticus network model calculates user by nervus opticus network model and puts question to corresponding text
Vector.User puts question to corresponding text vector to be referred to as question vector.
Step 210, the corresponding text vector of sample problem is obtained, puts question to user to corresponding text vector and sample problem
Corresponding text vector is matched, and is filtered out in sample problem and is putd question to matched multiple recommendation problems and recommendation to answer with user
Case.
Step 212, multiple recommendation problems and recommendation answer are sent to corresponding doctor terminal.
Problem base is set up on server, includes department, sample problem and corresponding sample answer in problem base.
Each department can correspond to multiple sample problems.Each sample problem has corresponding sample answer.
Server can obtain multiple sample problems in problem base, calculate each sample by nervus opticus network model
The corresponding text vector of problem.Wherein, the corresponding text vector of sample problem is referred to as sample vector.In order to effectively improve
Matching efficiency, server can first pass through in advance the sample of each sample problem in nervus opticus network model computational problem library to
Amount.
Server matches question vector with multiple sample vectors, obtains puing question to the multiple samples being close with user
Problem.Server will put question to the multiple sample problems being close that can mark as problem with user.Server can be from
Screening obtains the sample problem recommended Xiang doctor in multiple problems to be recommended, i.e. recommendation problem.
Server obtains the recommendation problem and corresponding recommendation answer of preset quantity, will recommend problem and recommends answer hair
It send to doctor terminal.Doctor by doctor terminal select in multiple recommendation problems with user's immediate recommendation problem of enquirement with
And recommend answer, the recommendation answer selected is sent to the user terminal.During making online interrogation, without doctor
Answer is thought deeply and be manually entered to raw put question to for user, can select most to connect with user's enquirement directly from recommendation problem
Close recommendation problem recommends answer to be sent to the user terminal for corresponding.
In the present embodiment, when user needs online interrogation, can be requested according to interrogation in the user identifier that carries to use
Corresponding doctor is distributed at family, establishes the communication connection between user terminal and doctor terminal, and whole to user terminal and doctor
The communication information between end is monitored.When listening to user's enquirement in the communication information, calculates user and put question to corresponding text
User is putd question to corresponding text vector text vector corresponding with sample problem to match, sieved in sample problem by vector
It selects and puts question to matched multiple recommendation problems with user and recommend answer, by multiple recommendation problems and answer is recommended to be sent to correspondence
Doctor terminal.During making online interrogation, think deeply without doctor for user's enquirement and defeated manually
Enter answer, can select to put question to immediate recommendation problem with user directly from recommendation problem, answer will be recommended to send out accordingly
It send to user terminal.To effectively shorten the time-consuming for replying user and puing question to.Due to recommendation problem be by carry out text to
Flux matched to obtain, recommendation problem is putd question to user and is close, and is putd question to using recommending answer to reply user, is effectively increased answer
Accuracy.
In one embodiment, as shown in figure 3, being that the step of user identifier distributes corresponding doctor has according to interrogation request
Body includes：
Step 302, a variety of user informations are obtained according to user identifier；
Step 304, vectorization processing is carried out to user information, obtains multi-C vector matrix；
Step 306, multi-C vector matrix is based on by first nerves network model and carries out prediction operation, obtain marking with user
Know corresponding department；
Step 308, corresponding doctor is distributed to user identifier according to department.
When server receives the interrogation request of user terminal uploads, server according to interrogation request in the user identifier that carries
A variety of user informations are obtained, including：Main suit's information, essential information, historical information etc..When server receives interrogation request, to
User terminal returns to corresponding problem.User can answer these problems by user terminal.User when answering these problems,
Content relevant to current symptomatic can be described.The answer of user is uploaded to server by user terminal, and server is marked
Be chief complaint information.Server can also collect corresponding basic letter while collection user identifier corresponding main suit's information parallel
Breath and historical information.When essential information can be user's registration application program, by the information of user terminal uploads to server,
Including：Age, gender etc..The user of all ages and classes or different sexes, it is possible to which there are similarities for main suit's information, but are asking
It needs to distribute to different departments when examining.Historical information include user by application program carry out online interrogation historical record,
The diagnosis and therapy recording etc. of present diagnosis and treatment mechanism.Identical department when would generally enter with first visit when due to user's further consultation, if user
This online interrogation is further consultation, then will be higher into the probability of same department.Therefore the historical information of user can aid in
Improve the accuracy of department's distribution.
Server carries out vector to a variety of user informations for being collected into and converts, and obtains and main suit information, essential information
And the corresponding multi-C vector matrix of historical information branch.Server merges multiple multi-C vector matrixes, generates first
The corresponding input vector matrix of neural network model.Server calls first nerves network model, using input vector matrix as
The input of first nerves network model, first nerves network model carry out operation, predict the corresponding department of multi-C vector matrix
Probability.Server can using the highest department of probability as the corresponding department of user identifier, and by the department be determined as to
The department of family mark distribution.It obtains server probability sorts from high to low, preset quantity (such as 2 or 3) department is determined
For the department distributed to user identifier.Server selects corresponding doctor according to the department of distribution, and doctor is distributed to user.
In the present embodiment, when user needs online interrogation, without manually selecting department, pass through first nerves network model
Prediction operation is carried out based on multi-C vector matrix, automatically and accurately department can be distributed for user, so as to be every use
The doctor that family distribution is adapted with its symptom.
In one embodiment, calculating the corresponding text vector of user's enquirement includes：User is putd question to and is pre-processed, is obtained
To pretreated text；Nervus opticus network model is called, nervus opticus network model includes LSTM and full articulamentum；It is logical
It crosses LSTM and operation is carried out to pretreated text, obtain characterization vector；Characterization vector is converted by full articulamentum, is obtained
To text vector corresponding with user's enquirement.
Server is that user distributes after corresponding department and corresponding doctor, and server establishes user terminal and doctor
Communication connection between terminal.Doctor can send problem to user terminal by doctor terminal, to further appreciate that user
Symptom.User can also be putd question to by user terminal Xiang doctor.
Server is putd question to user and is pre-processed.Pretreatment includes：Segment, go stop words, convoluted etc..Wherein, divide
Word, which refers to, is divided into participle to a sentence.It goes stop words to refer to and removes " grace " " " etc to the meaningless word of semantic understanding
Language.Convoluted refers to that the user if it is Hong Kong, Macao, Taiwan inputs using Chinese-traditional, then the use being inputted
Family information is converted to simplified form of Chinese Character.
Nervus opticus network model is pre-established on server, nervus opticus network model can be the problem of being in advance based on
What library was obtained by deep learning.After pretreatment is completed, server calls nervus opticus network model passes through nervus opticus
Network model puts question to the user after pretreatment and carries out vectorization processing.
When being communicated between user terminal and doctor terminal, the content that user puts question to may be with time gap before
Very long content is related.That is, the content of enquirement needs that context is combined to be identified when user puts question to.Based on this,
Two neural network models are taken using LSTM (Long Short-Term Memory, shot and long term memory network) and full articulamentum
It builds.Whether LSTM can identify two sentences of user terminal input semantically related.It is to accurately identify two sentences
Between the no context for there are semantically related, needing to input some user terminals incoherent information and redundancy into
Row abandons.Accurate operation result in order to obtain, LSTM is also required to be updated the content that user puts question to, to predict to obtain
Accurate result.For example, the gender of user diagnosis and treatment are judged it is critically important.User gender information for filling in registration may be with
Meaning, then doctor is before the answer for providing profession, it is necessary to confirm again to gender.If only abandoned initially
Gender information only adds new gender information, can all adversely affect to output result.
LSTM includes forgeing door, input gate and out gate.Wherein, incoherent information and redundancy can be believed by forgeing door
Breath abandoned, also just determine last moment subject of question how many can remain into current time.Input gate can be right
User put question to content be updated, that is, determine current time input subject of question how many can save.Output
Door can control the current output value how many subject of question can be output to LSTM.
In traditional mode, LSTM passes through forgetting door first and determines which subject of question is dropped, that is, passes into silence.It is losing
Forget and then is updated by the content that input gate puts question to user.It is exported then through out gate.But it is this
Mode is exported as a result, accuracy rate is lower.It, can be in traditional approach in order to effectively improve the accuracy rate of output result
LSTM is optimized.
Carrying out operation to pretreated text by LSTM in one of the embodiments, includes：It is current when getting
When the pretreated text at moment, forgetting processing is carried out to the pretreated text of last moment by forgeing door；It is losing
When forgetting to carry out forgeing processing, the pretreated text that current time inputs is updated by input gate；Pass through output
The text that door obtains obtained text after forgetting processing and update carries out operation, obtain it is corresponding with user's enquirement characterize to
Amount.
It, can be simultaneously to last moment when inputting the subject of question at current time by being optimized to traditional LSTM
User put question to and forget and the user at current time enquirement is updated.Thus, it is possible to the forgetting processing to door is forgotten
And the update processing of input gate is comprehensively considered, and the accuracy rate of output result is effectively increased.To subsequent prediction sample
There has also been a degree of promotions for the accuracy of this problem.
Multiple parameters are configured in LSTM, including forgeing the weight matrix of door and the weight square of bias term, input gate
Battle array and bias term, the weight matrix of out gate and bias term etc..
When a large amount of user carries out online interrogation simultaneously, the operand that LSTM faces is huge.In order to effectively simplify
Operand improves processing speed, in the present embodiment, can also optimize to LSTM further progress.It can be the parameter of forgetting door
Identical parameter is set with out gate.The weight matrix for forgeing door is identical as the weight matrix of out gate, forgets the biasing of door
Item is identical as the bias term of out gate, can allow and forget door and out gate using the progress operation of same formula.Thus
Operand can effectively be simplified, improve processing speed.
LSTM in nervus opticus network can calculate corresponding average value to the output at multiple moment, put question to as user
Characterization vector.Characterization vector is converted by the full articulamentum in nervus opticus network model, to obtain and user
Put question to corresponding text vector.It is matched by the text vector putd question to user text vector corresponding with sample text,
It can obtain problem to be recommended accordingly.Since the treatment effeciency of text vector is improved, matching effect is also promoted
The promotion of rate, it is quick Xiang doctor's progress sample problem recommendation so as to realize.
In one embodiment, as shown in figure 4, puing question to corresponding text vector text corresponding with sample problem user
Vector carries out matching：
Step 402, user is calculated to put question between corresponding text vector text vector corresponding with multiple sample problems
Similarity；
Step 404, when similarity reaches threshold value, the corresponding sample problem of similarity is labeled as problem to be recommended；
Step 406, it is screened in multiple problems to be recommended, by the problem flag to be recommended of the preset quantity filtered out
For recommendation problem corresponding with user's enquirement.
User puts question to corresponding text vector to be properly termed as question vector, and the corresponding text vector of sample problem is properly termed as
Sample vector.Server can put question to user and calculate corresponding question vector after receiving interrogation request, and to asking
Multiple sample problems in exam pool calculate corresponding sample vector.Server can also be in advance to sample problems all in problem base
Vectorization processing is carried out, in the database by the storage of corresponding sample vector.
Server is calculating the similarity between question vector and multiple sample vectors, is mentioned by similarity to user with this
It asks and is matched with sample problem.Wherein, similarity can be cosine similarity.When the phase between question vector and sample vector
When reaching threshold value like degree, indicating that user puts question to may be close with sample problem.Similarity can be more than threshold value by server
Sample problem corresponding to sample vector is labeled as problem to be recommended.Problem to be recommended can have multiple, different users to put question to
The quantity of corresponding problem to be recommended may be the same or different.Server can according to similarity to multiple to be recommended
Problem is screened, for example, can be ranked up according to the sequence of similarity from high to low, is screened since sequence is highest,
Obtain the problem to be recommended of preset quantity.What screening obtained is referred to as recommendation problem.
Server obtains the recommendation problem and corresponding recommendation answer of preset quantity, will recommend problem and recommends answer hair
It send to doctor terminal.Doctor terminal selects to put question to immediate recommendation problem with user in multiple recommendation problems and recommends to answer
The recommendation answer selected is sent to the user terminal by case.During making online interrogation, without doctor for use
Family, which is putd question to, is thought deeply and is manually entered answer, can select to put question to immediate recommendation with user directly from recommendation problem
Problem recommends answer to be sent to the user terminal for corresponding.
Matched efficiency is carried out in order to effectively improve user's enquirement with sample problem, it can be to the operand in matching process
Simplified.Specifically, server after the interrogation request for receiving user terminal uploads, can distribute corresponding for user
Department then selects corresponding doctor to distribute to user in department.When server listen to user terminal and doctor terminal it
Between the communication information in there are when user's enquirement, server can put question to addition department's information for user.To user put question into
Row pre-processes, and contains corresponding department's information in pretreated text.When server calls nervus opticus network model pair
When user puts question to progress vectorization processing, it can be and vectorization processing is carried out to the pretreated text comprising department's information.
It is all contained in thus obtained characterization vector and text vector and department's dimension.
When server, which puts question to user, to be matched with sample problem, since user puts question to corresponding text vector (i.e.
Question vector) in contain department's dimension, therefore server can select sample problem in identical department according to department's dimension
Corresponding sample vector is matched.To be effectively simplified operand, significantly in the case where not needing to carry out dimension-reduction treatment
User is improved to put question to and the matched efficiency of sample problem progress.
It should be understood that although each step in the flow chart of Fig. 2-4 is successively shown according to the instruction of arrow,
These steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly stating otherwise herein, these steps
Execution there is no stringent sequences to limit, these steps can execute in other order.Moreover, at least one in Fig. 2-4
Part steps may include that perhaps these sub-steps of multiple stages or stage are not necessarily in synchronization to multiple sub-steps
Completion is executed, but can be executed at different times, the execution sequence in these sub-steps or stage is also not necessarily successively
It carries out, but can be at least part of the sub-step or stage of other steps or other steps in turn or alternately
It executes.
In one embodiment, processing unit is replied based on the problem of online interrogation as shown in figure 5, providing one kind, wrapped
It includes：Communication module 502, monitors module 506, vectorization processing module 508 and matching module 510 at doctor's distribution module 504,
In：
Communication module 502, the interrogation for receiving user terminal uploads are requested；Interrogation request carries user identifier.
Doctor's distribution module 504, for being that user identifier distributes corresponding doctor according to interrogation request.
Communication module 502 is also used to establish the communication connection between user terminal and doctor terminal.
Module 506 is monitored, for the communication information between monitoring users terminal and doctor terminal.
When puing question to for listening to user in the communication information, it is corresponding to calculate user's enquirement for vectorization processing module 508
Text vector.
Matching module 510, for obtaining the corresponding text vector of sample problem, by user put question to corresponding text vector with
The corresponding text vector of sample problem is matched, and is filtered out in sample problem and is putd question to matched multiple recommendation problems with user
With recommendation answer.
Communication module 502 is also used to multiple recommendation problems and answer is recommended to be sent to corresponding doctor terminal.
In one embodiment, doctor's distribution module 504 is also used to obtain a variety of user informations according to user identifier；To with
Family information carries out vectorization processing, obtains multi-C vector matrix；By first nerves network model be based on multi-C vector matrix into
Row prediction operation, obtains department corresponding with user identifier；Corresponding doctor is distributed to user identifier according to department.
In one embodiment, vectorization processing module 508 is also used to put question to user and pre-process, and is pre-processed
Text afterwards；Nervus opticus network model is called, nervus opticus network model includes LSTM and full articulamentum；Pass through LSTM pairs
Pretreated text carries out operation, obtains characterization vector；Characterization vector is converted by full articulamentum, is obtained and user
Put question to corresponding text vector.
In one embodiment, LSTM includes forgeing door, input gate and out gate, vectorization processing module 508 to be also used to
When getting the pretreated text at current time, the pretreated text of last moment is lost by forgeing door
Forget to handle；When forgetting door carries out forgeing processing, the pretreated text that current time inputs is carried out more by input gate
Newly；Operation is carried out by the text that out gate obtains the text and update that obtain after forgetting processing, obtains puing question to user
Corresponding characterization vector.
In one embodiment, matching module 510 is also used to calculate user and puts question to corresponding text vector and multiple samples
Similarity between the corresponding text vector of problem；When similarity reaches threshold value, the corresponding sample problem of similarity is marked
For problem to be recommended；It is screened in multiple problems to be recommended, is by the problem flag to be recommended of the preset quantity filtered out
Recommendation problem corresponding with user's enquirement.
The problem of about based on online interrogation reply processing unit it is specific restriction may refer to above for be based on
The problem of line interrogation, replies the restriction of processing method, and details are not described herein.Above-mentioned the problem of being based on online interrogation, replies processing dress
Modules in setting can be realized fully or partially through software, hardware and combinations thereof.Above-mentioned each module can be in the form of hardware
It is embedded in or independently of the storage that in the processor in computer equipment, can also be stored in a software form in computer equipment
In device, the corresponding operation of the above modules is executed in order to which processor calls.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 6.The computer equipment include by system bus connect processor, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data
Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The database of machine equipment is for storing the data such as sample vector.The network interface of the computer equipment is used for logical with external terminal
Cross network connection communication.To realize that a kind of the problem of being based on online interrogation answer is handled when the computer program is executed by processor
Method.
It will be understood by those skilled in the art that structure shown in Fig. 6, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specific computer equipment
It may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
In one embodiment, a kind of computer equipment, including memory and processor are provided, which is stored with
Computer program, the processor realize the step in above-mentioned each embodiment of the method when executing computer program.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program realizes the step in above-mentioned each embodiment of the method when being executed by processor.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. one kind is based on the problem of online interrogation and replies processing method, the method includes：
Receive the interrogation request of user terminal uploads；The interrogation request carries user identifier；
Be that the user identifier distributes corresponding doctor according to interrogation request, establish the user terminal and doctor terminal it
Between communication connection；
Monitor the communication information between the user terminal and the doctor terminal；
When listening to user's enquirement in the communication information, calculates the user and put question to corresponding text vector；
The corresponding text vector of sample problem is obtained, puts question to corresponding text vector corresponding with the sample problem user
Text vector matched, filtered out in the sample problem and put question to matched multiple recommendation problems with the user and push away
Recommend answer；
By multiple recommendation problems and answer is recommended to be sent to corresponding doctor terminal.
2. the method according to claim 1, wherein described divide according to interrogation request for the user identifier
Include with corresponding doctor：
A variety of user informations are obtained according to the user identifier；
Vectorization processing is carried out to the user information, obtains multi-C vector matrix；
The multi-C vector matrix is based on by first nerves network model and carries out prediction operation, is obtained and the user identifier pair
The department answered；
Corresponding doctor is distributed to the user identifier according to the department.
3. the method according to claim 1, wherein described calculate the corresponding text vector packet of user's enquirement
It includes：
The user is putd question to and is pre-processed, pretreated text is obtained；
Nervus opticus network model is called, the nervus opticus network model includes LSTM and full articulamentum；
Operation is carried out to pretreated text by the LSTM, obtains characterization vector；
The characterization vector is converted by the full articulamentum, obtains text vector corresponding with user enquirement.
4. according to the method described in claim 3, it is characterized in that, the LSTM includes forgeing door, input gate and out gate, institute
It states and includes to the progress operation of pretreated text by the LSTM：
When getting the pretreated text at current time, by forget door to the pretreated text of last moment into
Row forgetting processing；
When the forgetting door carries out forgeing processing, current time is inputted by the input gate pretreated text into
Row updates；
Operation is carried out by the text that out gate obtains the text and update that obtain after forgetting processing, obtains puing question to user
Corresponding characterization vector.
5. method according to claim 1-4, which is characterized in that described to put question to the user to corresponding text
Vector text vector corresponding with the sample problem carries out matching：
It calculates the user and puts question to similarity between corresponding text vector text vector corresponding with multiple sample problems；
When the similarity reaches threshold value, the corresponding sample problem of the similarity is labeled as problem to be recommended；
It is screened in multiple problems to be recommended, is and the user by the problem flag to be recommended of the preset quantity filtered out
Put question to corresponding recommendation problem.
6. one kind is based on the problem of online interrogation and replies processing unit, which is characterized in that described device includes：
Communication module, the interrogation for receiving user terminal uploads are requested；The interrogation request carries user identifier；
Doctor's distribution module, for being that the user identifier distributes corresponding doctor according to interrogation request；
The communication module is also used to establish the communication connection between the user terminal and doctor terminal；
Module is monitored, for monitoring the communication information between the user terminal and the doctor terminal；
Vectorization processing module when puing question to for listening to user in the communication information, calculates the user and puts question to correspondence
Text vector；
Matching module puts question to the user to corresponding text vector and institute for obtaining the corresponding text vector of sample problem
It states the corresponding text vector of sample problem to be matched, be filtered out in the sample problem matched more with user enquirement
A recommendation problem and recommendation answer；
The communication module is also used to multiple recommendation problems and answer is recommended to be sent to corresponding doctor terminal.
7. device according to claim 6, which is characterized in that doctor's distribution module is also used to be marked according to the user
Know and obtains a variety of user informations；Vectorization processing is carried out to the user information, obtains multi-C vector matrix；Pass through first nerves
Network model is based on the multi-C vector matrix and carries out prediction operation, obtains department corresponding with the user identifier；According to institute
It states department and distributes corresponding doctor to the user identifier.
8. device according to claim 6, which is characterized in that the vectorization processing module is also used to mention the user
It asks and is pre-processed, obtain pretreated text；Nervus opticus network model is called, the nervus opticus network model includes
LSTM and full articulamentum；Operation is carried out to pretreated text by the LSTM, obtains characterization vector；By described complete
Articulamentum converts the characterization vector, obtains text vector corresponding with user enquirement.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 7 the method when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810606392.4A CN108922634A (en) | 2018-06-13 | 2018-06-13 | The problem of based on online interrogation, replies processing method, device and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810606392.4A CN108922634A (en) | 2018-06-13 | 2018-06-13 | The problem of based on online interrogation, replies processing method, device and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108922634A true CN108922634A (en) | 2018-11-30 |
Family
ID=64420525
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810606392.4A Pending CN108922634A (en) | 2018-06-13 | 2018-06-13 | The problem of based on online interrogation, replies processing method, device and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108922634A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109635091A (en) * | 2018-12-14 | 2019-04-16 | 上海钛米机器人科技有限公司 | A kind of method for recognizing semantics, device, terminal device and storage medium |
| CN112035674A (en) * | 2020-08-28 | 2020-12-04 | 康键信息技术（深圳）有限公司 | Diagnosis guide data acquisition method and device, computer equipment and storage medium |
| CN112699224A (en) * | 2019-10-22 | 2021-04-23 | 中信百信银行股份有限公司 | Question-answer dialogue method and device, electronic equipment and computer readable storage medium |
| CN112786176A (en) * | 2021-02-22 | 2021-05-11 | 北京融威众邦电子技术有限公司 | Intelligent self-service diagnosis method and device and computer equipment |
| CN113192500A (en) * | 2020-01-29 | 2021-07-30 | 丰田自动车株式会社 | Proxy device, proxy system, and non-transitory recording medium |
| US11475068B2 (en) * | 2019-07-26 | 2022-10-18 | Beijing Boe Technology Development Co., Ltd. | Automatic question answering method and apparatus, storage medium and server |
| US20230051306A1 (en) * | 2021-08-11 | 2023-02-16 | International Business Machines Corporation | Multi-user voice assistant with disambiguation |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105653840A (en) * | 2015-12-21 | 2016-06-08 | 青岛中科慧康科技有限公司 | Similar case recommendation system based on word and phrase distributed representation, and corresponding method |
| CN106649561A (en) * | 2016-11-10 | 2017-05-10 | 复旦大学 | Intelligent question-answering system for tax consultation service |
| CN106650789A (en) * | 2016-11-16 | 2017-05-10 | 同济大学 | Image description generation method based on depth LSTM network |
| CN106650261A (en) * | 2016-12-22 | 2017-05-10 | 上海智臻智能网络科技股份有限公司 | Intelligent inquiry method, device and system |
| CN107247868A (en) * | 2017-05-18 | 2017-10-13 | 深思考人工智能机器人科技（北京）有限公司 | A kind of artificial intelligence aids in interrogation system |
| CN108062303A (en) * | 2017-12-06 | 2018-05-22 | 北京奇虎科技有限公司 | The recognition methods of refuse messages and device |
-
2018
- 2018-06-13 CN CN201810606392.4A patent/CN108922634A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105653840A (en) * | 2015-12-21 | 2016-06-08 | 青岛中科慧康科技有限公司 | Similar case recommendation system based on word and phrase distributed representation, and corresponding method |
| CN106649561A (en) * | 2016-11-10 | 2017-05-10 | 复旦大学 | Intelligent question-answering system for tax consultation service |
| CN106650789A (en) * | 2016-11-16 | 2017-05-10 | 同济大学 | Image description generation method based on depth LSTM network |
| CN106650261A (en) * | 2016-12-22 | 2017-05-10 | 上海智臻智能网络科技股份有限公司 | Intelligent inquiry method, device and system |
| CN107247868A (en) * | 2017-05-18 | 2017-10-13 | 深思考人工智能机器人科技（北京）有限公司 | A kind of artificial intelligence aids in interrogation system |
| CN108062303A (en) * | 2017-12-06 | 2018-05-22 | 北京奇虎科技有限公司 | The recognition methods of refuse messages and device |
Non-Patent Citations (2)
| Title |
|---|
| 包晓安 等: ""基于LSTM的智能家居机器学习系统预测模型研究"", 《浙江理工大学学报(自然科学版)》, vol. 39, no. 02 * |
| 包晓安 等: ""基于LSTM的智能家居机器学习系统预测模型研究"", 《浙江理工大学学报（自然科学版）》, vol. 39, no. 2 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109635091A (en) * | 2018-12-14 | 2019-04-16 | 上海钛米机器人科技有限公司 | A kind of method for recognizing semantics, device, terminal device and storage medium |
| US11475068B2 (en) * | 2019-07-26 | 2022-10-18 | Beijing Boe Technology Development Co., Ltd. | Automatic question answering method and apparatus, storage medium and server |
| CN112699224A (en) * | 2019-10-22 | 2021-04-23 | 中信百信银行股份有限公司 | Question-answer dialogue method and device, electronic equipment and computer readable storage medium |
| CN112699224B (en) * | 2019-10-22 | 2024-05-17 | 中信百信银行股份有限公司 | Question-answering dialogue method, question-answering dialogue device, electronic equipment and computer readable storage medium |
| CN113192500A (en) * | 2020-01-29 | 2021-07-30 | 丰田自动车株式会社 | Proxy device, proxy system, and non-transitory recording medium |
| CN112035674A (en) * | 2020-08-28 | 2020-12-04 | 康键信息技术（深圳）有限公司 | Diagnosis guide data acquisition method and device, computer equipment and storage medium |
| CN112035674B (en) * | 2020-08-28 | 2023-05-23 | 康键信息技术（深圳）有限公司 | Diagnosis guiding data acquisition method, device, computer equipment and storage medium |
| CN112786176A (en) * | 2021-02-22 | 2021-05-11 | 北京融威众邦电子技术有限公司 | Intelligent self-service diagnosis method and device and computer equipment |
| US20230051306A1 (en) * | 2021-08-11 | 2023-02-16 | International Business Machines Corporation | Multi-user voice assistant with disambiguation |
| US11830490B2 (en) * | 2021-08-11 | 2023-11-28 | International Business Machines Corporation | Multi-user voice assistant with disambiguation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108922634A (en) | The problem of based on online interrogation, replies processing method, device and computer equipment | |
| CN109065139A (en) | Medical follow up method, apparatus, computer equipment and storage medium | |
| CN112035674B (en) | Diagnosis guiding data acquisition method, device, computer equipment and storage medium | |
| CN108986910B (en) | On-line question and answer method, device, computer equipment and storage medium | |
| US11250951B2 (en) | Feature engineering method, apparatus, and system | |
| WO2021151327A1 (en) | Triage data processing method and apparatus, and device and medium | |
| CN109816220A (en) | Quality of service monitoring and treating method and apparatus based on intelligent decision | |
| CN112035611A (en) | Target user recommendation method and device, computer equipment and storage medium | |
| CN115238173B (en) | Behavior analysis and medical service pushing method, equipment and medium based on big data | |
| CN113112282A (en) | Method, device, equipment and medium for processing consult problem based on client portrait | |
| WO2021151356A1 (en) | Triage data processing method and apparatus, and computer device and storage medium | |
| CN111415760B (en) | Doctor recommendation method, doctor recommendation system, computer equipment and storage medium | |
| CN114974501A (en) | Artificial intelligence based medicine recommendation method, device, equipment and storage medium | |
| CN113724815A (en) | Information pushing method and device based on decision grouping model | |
| CN112925888A (en) | Method and device for training question-answer response and small sample text matching model | |
| CN112989179B (en) | Model training and multimedia content recommendation method and device | |
| CN112397197A (en) | Artificial intelligence-based inquiry data processing method and device | |
| CN114860887A (en) | Disease content pushing method, device, equipment and medium based on intelligent association | |
| CN112036161A (en) | Requirement document processing method, device, equipment and storage medium | |
| CN112330400B (en) | Product recommendation method and device, computer equipment and storage medium | |
| CN109003193B (en) | Method, device, computer equipment and storage medium for predicting insurance risk | |
| CN116894498A (en) | Training method, strategy identification method, device and equipment of network model | |
| CN112925723B (en) | Test service recommendation method and device, computer equipment and storage medium | |
| CN116992861B (en) | Intelligent medical service processing method and system based on data processing | |
| CN113239154B (en) | Quality inspection method and device for seat conversation operation, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |