CN109065015A - A kind of collecting method, device, equipment and readable storage medium storing program for executing - Google Patents
A kind of collecting method, device, equipment and readable storage medium storing program for executing Download PDFInfo
- Publication number
- CN109065015A CN109065015A CN201810844009.9A CN201810844009A CN109065015A CN 109065015 A CN109065015 A CN 109065015A CN 201810844009 A CN201810844009 A CN 201810844009A CN 109065015 A CN109065015 A CN 109065015A
- Authority
- CN
- China
- Prior art keywords
- answer
- question
- node
- information
- answer node
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000012549 training Methods 0.000 claims description 32
- 201000010099 disease Diseases 0.000 claims description 22
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 22
- 238000013518 transcription Methods 0.000 claims description 16
- 230000035897 transcription Effects 0.000 claims description 16
- 238000000605 extraction Methods 0.000 claims description 14
- 208000024891 symptom Diseases 0.000 claims description 12
- 238000012163 sequencing technique Methods 0.000 claims description 9
- 239000003814 drug Substances 0.000 claims description 7
- 239000000284 extract Substances 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 238000000151 deposition Methods 0.000 claims 2
- 230000015572 biosynthetic process Effects 0.000 description 7
- 238000004891 communication Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 7
- 238000003786 synthesis reaction Methods 0.000 description 7
- 238000003745 diagnosis Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 239000000463 material Substances 0.000 description 3
- 208000004998 Abdominal Pain Diseases 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012015 optical character recognition Methods 0.000 description 2
- 206010019233 Headaches Diseases 0.000 description 1
- 206010037660 Pyrexia Diseases 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011049 filling Methods 0.000 description 1
- 231100000869 headache Toxicity 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000001356 surgical procedure Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
- G10L13/02—Methods for producing synthetic speech; Speech synthesisers
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
- G10L15/063—Training
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
- G10L15/063—Training
- G10L2015/0638—Interactive procedures
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L2015/225—Feedback of the input speech
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
This application discloses a kind of collecting method, device, equipment and readable storage medium storing program for executing, due to obtaining the corresponding question and answer node set of destination item, the corresponding problem information of destination item is contained in set, the machine automatic data acquisition realized based on problem information, the problem of being not in acquisition shortage of data caused by artificial leakage is asked, and machine acquisition is greatly promoted compared to artificial collecting efficiency.
Description
Technical field
This application involves natural language processing technique field, more specifically to a kind of collecting method, device,
Equipment and readable storage medium storing program for executing.
Background technique
With the development of the times, data age has currently been come into.All trades and professions require accumulation basic data, with branch
Hold higher level decision.
Such as, hearing content-data acquisition needs to acquire by question and answer mode by the answer of hearing people more typical example, and
Ultimately form hearing record.Hearing record can be used as the support material of subsequent case trial.For another example, case acquires, i.e., logical
The question and answer exchange between doctors and patients is crossed, morbidity is obtained by doctor and treatment is passed through, and forms case book.Case acquisition is diagnosis disease
One of the important evidence of disease, the support material as medical diagnosis on disease.
By the study found that the data acquisition of the question and answer mode of existing projects be by manually realizing, i.e., by
Enquirement side proposes problem, by answer side to corresponding answer of ging wrong, and by the manually recorded problem in enquirement side and corresponding answer
Content.Obviously, this data acquisition modes are influenced by enquirement side personal experience and state, for complicated project, it is easy to
It is incomplete to there is the consideration of enquirement side, leads to the problem of acquiring shortage of data.Also, artificial acquisition is asked there is also collecting efficiency is low
Topic.
Summary of the invention
In view of this, this application provides a kind of collecting method, device, equipment and readable storage medium storing program for executing, for solving
The problem of acquisition data present in certainly existing hand data collection easily lack, are at high cost, low efficiency.
To achieve the goals above, it is proposed that scheme it is as follows:
A kind of collecting method, comprising:
Obtain the corresponding question and answer node set of destination item of pending data acquisition, the question and answer node set include with
The corresponding question and answer node of the destination item, the question and answer node include problem information;
Question and answer node is chosen from the question and answer node set, and letter the problem of the question and answer node that exports selection is included
Breath;
The answer information to output problem information feedback is obtained, the corresponding answer information of question and answer node is obtained.
It is preferably, described that question and answer node is chosen from the question and answer node set, comprising:
According to the inquiry sequence of the corresponding each question and answer node of the preset destination item, from the question and answer node set
Choose question and answer node.
Preferably, the inquiry sequence according to the corresponding each question and answer node of the preset destination item, is asked from described
Answer selection question and answer node in node set, comprising:
According to the collating sequence of each question and answer node in the question and answer node set, question and answer node is from first to last chosen, it is described
The collating sequence of each question and answer node and the inquiry sequence consensus in question and answer node set.
Preferably, the question and answer node also includes next question and answer node slot, is determined for storing according to the inquiry sequence
The question and answer node next question and answer node index;
The inquiry sequence according to the corresponding each question and answer node of the preset destination item, from the question and answer node collection
Question and answer node is chosen in conjunction, comprising:
It, will when determination needs to choose next question and answer node in the corresponding answer information of question and answer node currently chosen
The index of the next question and answer node for next question and answer node slot storage that the question and answer node currently chosen is included is corresponding to ask
Node is answered, as next question and answer node.
It is preferably, described that question and answer node is chosen from the question and answer node set, comprising:
For the problem that each question and answer node chosen in the question and answer node set, according to the question and answer node information
And information is answered, determine the node diagnostic of the question and answer node;
According to the sequencing of selection, the node diagnostic group for each question and answer node chosen is combined into node diagnostic collection
It closes;
The node diagnostic set is inputted to preset node preference pattern, obtains next the asking of node preference pattern output
Answer the index of node;
The node preference pattern is, with the corresponding node diagnostic training number for having chosen question and answer node of the destination item
It is combined into training sample according to the node diagnostic training dataset being combined into according to selection sequence, with next question and answer node to be chosen of mark
Index be sample label training obtain.
Preferably, the question and answer node also includes next question and answer node slot, for storing the index of next question and answer node;
It is described that question and answer node is chosen from the question and answer node set, further includes:
Sentence when determination needs to choose next question and answer node in the question and answer node corresponding answer information currently chosen
Break and whether is stored with the index of next question and answer node in next question and answer node slot that the question and answer node currently chosen is included;
If so, the next question and answer section for the next question and answer node slot storage for being included by the question and answer node currently chosen
The corresponding question and answer node of index of point, as next question and answer node;
If it is not, then executing each question and answer node for being directed to and having chosen in the question and answer node set, asked according to described
The problem of answering node information and answer information, determine the operation of the node diagnostic of the question and answer node.
Preferably, it is described according to information the problem of the question and answer node and answer information, determine the section of the question and answer node
Point feature, comprising:
Using information the problem of the question and answer node and information is answered as input data, inputs preset nodes encoding mould
Type, the nodes encoding model are that feature extraction, and the feature according to extraction can be carried out to input data, predict third party
The model of the project result of project, third party's project are the project using destination item data collected;
The feature that the nodes encoding model extracts the input data is obtained, the node as the question and answer node is special
Sign.
Preferably, the problem of question and answer node that the output is chosen is included information, comprising:
If described problem information be textual form, export in the form of text selection question and answer node it is included the problem of letter
Breath, or, the problem of question and answer node by selection is included information progress speech synthesis, and the problem of export the speech form of synthesis
Information;
If described problem information is speech form, the problem of question and answer node chosen is included letter is exported with speech form
Breath or, the problem of question and answer node by selection is included information carries out speech transcription, and exports asking for the textual form after transcription
Inscribe information.
Preferably, the answer information for obtaining information feedback the problem of to output, obtains the corresponding answer of question and answer node
Information, comprising:
The answer information of the speech form of information feedback the problem of to output is obtained, and is returning for textual form by its transcription
Answer information;Or,
The answer information of the image format of information feedback the problem of to output is obtained, and image text identification is carried out to it,
Identify the answer information of textual form;Or,
Obtain the answer information of the textual form of information feedback the problem of to output;
The answer information of acquisition is standardized, the corresponding standard of question and answer node is obtained and answers information.
Preferably, the question and answer node also includes problem types slot, the type for storage problem information;
The answer information of described pair of acquisition is standardized, and is obtained the corresponding standard of question and answer node and is answered information, comprising:
If according to problem types slot determine obtain answers information correspondence problem information type for whether class problem, root
It to class certainly or negate class keywords comprising situation according to the answer information of acquisition, it is certainly or no that the standard that determines, which answers information,
It is fixed;
If the type for determining the answer information correspondence problem information obtained according to problem types slot is description class problem, will
The answer information of acquisition answers information as standard.
Preferably, the question and answer node also includes candidate answers slot, for storing and the matched candidate answers of problem information
Information;
The answer information of described pair of acquisition is standardized, and is obtained the corresponding standard of question and answer node and is answered information, further includes:
If the type for determining the answer information correspondence problem information obtained according to problem types slot is selection class problem, count
Calculate the similarity of the answer information obtained and each candidate answers information stored in candidate answers slot;
According to the size of similarity, determine that standard answers information from candidate answers information.
Preferably, the destination item includes case acquisition project, hearing content acquisition project, interview data acquisition projects
In any one or more.
Preferably, the destination item is that case acquires project, then the destination item corresponding question and answer node set
Generating process, comprising:
The corresponding department's disease of project is acquired according to case, obtains symptom terms relevant to department's disease;
Career in medicine knowledge, which is answered, collects question and answer data relevant to the symptom terms in resource, and is organized into problem information and returns
Answer information;
The problem information after node is formed by the problem information node after arrangement, and according to preset interrogation process
Question and answer node set.
A kind of data acquisition device, comprising:
Question and answer node set acquiring unit, for obtaining the corresponding question and answer node collection of destination item of pending data acquisition
It closes, the question and answer node set includes question and answer node corresponding with the destination item, and the question and answer node includes problem information;
Question and answer node selection unit, for choosing question and answer node from the question and answer node set;
Problem information output unit, for exporting the problem of question and answer node chosen is included information;
Information acquisition unit is answered, for obtaining the answer information to information feedback the problem of output, obtains question and answer node
Corresponding answer information.
Preferably, the question and answer node selection unit includes:
Sequentially selection unit, for the inquiry sequence according to the corresponding each question and answer node of the preset destination item, from
Question and answer node is chosen in the question and answer node set.
Preferably, the sequentially selection unit includes:
Sequence selection unit in gathering, for the collating sequence according to each question and answer node in the question and answer node set, from
Head to tail chooses question and answer node, the collating sequence of each question and answer node and the inquiry sequence consensus in the question and answer node set.
Preferably, the question and answer node also includes next question and answer node slot, is determined for storing according to the inquiry sequence
The question and answer node next question and answer node index;The sequentially selection unit includes:
According to index selection unit, in the corresponding answer information of question and answer node currently chosen, determination to need to select
When removing a question and answer node, the next question and answer for the next question and answer node slot storage for being included by the question and answer node currently chosen
The corresponding question and answer node of the index of node, as next question and answer node.
Preferably, the question and answer node selection unit includes:
Node diagnostic determination unit, for being directed to each question and answer node chosen in the question and answer node set, according to
The problem of question and answer node information and answer information, determine the node diagnostic of the question and answer node;
Feature assembled unit, for the sequencing according to selection, by the node diagnostic for each question and answer node chosen
Group is combined into node diagnostic set;
Node preference pattern predicting unit is obtained for the node diagnostic set to be inputted to preset node preference pattern
The index of the next question and answer node exported to node preference pattern;
The node preference pattern is, with the corresponding node diagnostic training number for having chosen question and answer node of the destination item
It is combined into training sample according to the node diagnostic training dataset being combined into according to selection sequence, with next question and answer node to be chosen of mark
Index be sample label training obtain.
Preferably, the question and answer node also includes next question and answer node slot, for storing the index of next question and answer node;
The question and answer node selection unit further include:
Question and answer node slot judging unit, in the corresponding answer information of question and answer node currently chosen, determining to be needed
When choosing next question and answer node, judge whether deposit in next question and answer node slot that the question and answer node currently chosen is included
Contain the index of next question and answer node;If so, executing question and answer node slot uses unit, determined if it is not, executing the node diagnostic
Unit;
The question and answer node slot uses unit, next question and answer section for being included by the question and answer node currently chosen
The corresponding question and answer node of index of next question and answer node of point slot storage, as next question and answer node.
Preferably, the node diagnostic determination unit includes:
Nodes encoding model prediction unit, for using information the problem of the question and answer node and answer information as input number
According to, preset nodes encoding model is inputted, the nodes encoding model is, it can carry out feature extraction to input data, and according to
According to the feature of extraction, the model of the project result of third party's project is predicted, third party's project is to apply the destination item
The project of data collected;
Nodes encoding aspect of model extraction unit extracts the input data for obtaining the nodes encoding model
Feature, the node diagnostic as the question and answer node.
Preferably, described problem information output unit includes:
First problem information exports subelement, if being textual form for described problem information, exports in the form of text
The problem of question and answer node of selection is included information, or, the problem of question and answer node by selection is included information carries out voice conjunction
At, and information the problem of export the speech form of synthesis;
Second Problem information exports subelement, if being speech form for described problem information, is exported with speech form
The problem of question and answer node of selection is included information turns or, the problem of question and answer node by selection is included information carries out voice
The problem of writing, and exporting the textual form after transcription information.
Preferably, the answer information acquisition unit includes:
Voice answering acquisition of information subelement, for obtaining the answer letter to the speech form of information feedback the problem of output
Breath, and be the answer information of textual form by its transcription;Or,
Image answers acquisition of information subelement, for obtaining the answer letter to the image format of information feedback the problem of output
Breath, and image text identification is carried out to it, identify the answer information of textual form;Or,
Text answers acquisition of information subelement, for obtaining the answer letter to the textual form of information feedback the problem of output
Breath;
Standardization unit obtains the corresponding standard of question and answer node for being standardized to the answer information of acquisition
Answer information.
Preferably, the question and answer node also includes problem types slot, the type for storage problem information;
The standardization unit includes:
First standardization subelement, if for determining that the answer information correspondence problem obtained is believed according to problem types slot
The type of breath be whether class problem, then to class certainly or negate class keywords comprising situation according to the answers information of acquisition, really
The quasi- information of answering of calibration is positive or negative;
Second standardization subelement, if for determining that the answer information correspondence problem obtained is believed according to problem types slot
The type of breath is description class problem, then the answer information that will acquire answers information as standard.
Preferably, the question and answer node also includes candidate answers slot, for storing and the matched candidate answers of problem information
Information;
The standardization unit further include:
Third standardization subelement, if for determining that the answer information correspondence problem obtained is believed according to problem types slot
The type of breath is selection class problem, then calculates each candidate answers information stored in the answer information and candidate answers slot of acquisition
Similarity;According to the size of similarity, determine that standard answers information from candidate answers information.
A kind of data acquisition equipment, including memory and processor;
The memory, for storing program;
The processor realizes each step of collecting method as described above for executing described program.
A kind of readable storage medium storing program for executing is stored thereon with computer program, real when the computer program is executed by processor
Now each step of collecting method as described above.
It can be seen from the above technical scheme that collecting method provided by the embodiments of the present application, obtains destination item
Corresponding question and answer node set, which contains question and answer node corresponding with destination item, and question and answer node is believed comprising problem
Breath, and therefrom choose question and answer node, export selection question and answer node it is included the problem of information, for user be directed to problem information it is anti-
It is fed back to and answers information, and get the answer information of the feedback, obtain the corresponding answer information of question and answer node.The application is due to obtaining
To the corresponding question and answer node set of destination item, the corresponding problem information of destination item is contained in set, is based on problem information
The problem of machine automatic data acquisition of realization is not in acquisition shortage of data caused by artificial leakage is asked, and machine acquires
It is greatly promoted compared to artificial collecting efficiency.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of application for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is a kind of collecting method flow chart disclosed in the embodiment of the present application;
Fig. 2 is a kind of data acquisition device structural schematic diagram disclosed in the embodiment of the present application;
Fig. 3 is a kind of hardware block diagram of data acquisition equipment disclosed in the embodiment of the present application.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of embodiments of the present application, instead of all the embodiments.It is based on
Embodiment in the application, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall in the protection scope of this application.
Data acquisition plan provided by the embodiments of the present application, can be applied to data acquisition equipment, such as computer, mobile phone, clothes
The intelligent terminals such as business device, the intelligent terminal can carry out data interaction with user, and interactive mode be not limited to: voice,
A variety of interactive modes such as text, image.The data acquisition plan of the present embodiment can be directed to the target of any one question and answer mode
Project, such as case acquisition project, hearing content acquisition project, interview data acquisition projects.It should be noted that case acquisition
The purpose of project is to obtain case book, which is not medical diagnosis on disease as a result, but for assisting doctor to carry out medical diagnosis on disease
Support material.
Next, being introduced in conjunction with collecting method of the Fig. 1 to this embodiment of the present application, as shown in Figure 1, this method
May include:
Step S100, the corresponding question and answer node set of destination item of pending data acquisition is obtained.
Wherein, the question and answer node set includes question and answer node corresponding with the destination item, the question and answer node packet
Containing problem information.
As previously mentioned, destination item can be the project for needing to carry out data acquisition by question and answer mode.According to target item
Purpose is different, and corresponding question and answer node set is also different.The corresponding all question and answer of destination item are contained in question and answer node set
Node, each question and answer node include corresponding problem information, and problem information can be understood as the description information of problem, such as " whether
Appearance abdominal pain ", " duration or discontinuity pectoralgia " etc..The question and answer node for including in question and answer node set at least one, one
As property, the number of question and answer nodes is multiple.
By taking destination item is case acquisition project as an example, it can also be further subdivided into sub-project, such as according to medical
The difference of department can be divided into multiple and different department's case acquisition sub-projects.
The application is directed to different destination items in advance, it is determined that needs the problem of being acquired information, and is based on this structure
The corresponding question and answer node set of destination item is built.When carrying out the destination item of data acquisition required for having determined, Ke Yizhi
It obtains and takes pre-generated question and answer node set corresponding with the destination item.
Step S110, question and answer node is chosen from the question and answer node set, and the question and answer node for exporting selection is included
The problem of information.
Specifically, question and answer node can be chosen from question and answer node set, and the question and answer node for exporting the selection is included
The problem of information.Selection mode can be to be chosen one by one, is also possible to once choose multiple.
It is understood that the form of problem information can there are many, such as textual form, speech form.If problem is believed
The form of breath is textual form, then can directly export information the problem of textual form, is such as carried out by way of display screen
It shows, carries out disclosure for user.Further, it is also possible to problem information is carried out speech synthesis by textual form, and after exporting synthesis
Speech form the problem of information.Specifically, information the problem of speech form after synthesis can be broadcast by microphone
It puts, is listened to for user.
Further, if the form of problem information is speech form, problem information can be exported with speech form.In addition,
The problem of problem information can also being subjected to speech transcription, and exporting the textual form after transcription information.
Certainly above-mentioned to merely illustrate the output form of several problem informations, it in addition to this can also be right in other forms
Problem information is exported, and guarantees that user can recognize problem information.
Step S120, the answer information to output problem information feedback is obtained, the corresponding answer information of question and answer node is obtained.
Specifically, the problem of including to the question and answer node of selection in previous step information exports, and on this basis, uses
Family information feedback can answer information aiming at the problem that output.The answer of the problem of being obtained in this step to output information feedback is believed
Breath, the answer information and information the problem of output are corresponding namely corresponding with the question and answer node where problem information, therefore can obtain
To the corresponding answer information of question and answer node.
It is understood that the answer information obtained in this step can be voice answering information, it can also be that text returns
Information is answered, alternatively, can also be that the text of the other forms such as image answers information.By taking case acquisition project as an example, patient can be with
Information is answered by voice, textual form feedback, checklist can also will be checked as answer information.
Due to choosing question and answer node in previous step from question and answer node set, and the question and answer node for exporting selection includes
Problem information, therefore, by the corresponding answer information of question and answer node each in question and answer node set available in this step, most
The corresponding answer information of each question and answer node in question and answer node set is obtained eventually.The corresponding answer information of each question and answer node is target
The corresponding acquisition data of project.
Collecting method provided by the embodiments of the present application, due to getting the corresponding question and answer node set of destination item,
The corresponding problem information of destination item is contained in set, based on the machine automatic data acquisition that problem information is realized, will not be gone out
Caused by now artificial leakage is asked the problem of acquisition shortage of data, and machine acquisition has obtained mentioning significantly compared to artificial collecting efficiency
It rises.
In one embodiment of the application, to above-mentioned steps S120, the answer of information feedback the problem of to output is obtained
Information obtains the corresponding process for answering information of question and answer node and is illustrated.
Above-mentioned to have been described above, answer information can there are many forms, such as speech form, image format, textual form.
For the ease of editing to answer information, the various forms of answer information that the present embodiment will acquire are converted into text shape
Formula specifically includes:
If 1) answering information is speech form, the present embodiment obtains the speech form of information feedback the problem of to output
Information is answered, and is the answer information of textual form by its transcription.
Specifically, in order to improve speech transcription accuracy rate, corresponding voice training can be obtained in advance for destination item
Data, and the corresponding content of text of semantic training data is marked, and then instruct using voice training data and corresponding content of text
Practice speech transcription model.It is subsequent to can use trained speech transcription model, transcription is carried out to information is answered, is obtained corresponding
The answer information of textual form.
By taking destination item is case acquisition as an example, the true sufferer that voice training data can be collection is converged in treatment process
The voice data that always doctor's problem is answered.
If 2) answering information is image format, the present embodiment obtains the image format of information feedback the problem of to output
Information is answered, and image text identification is carried out to it, identifies the answer information of textual form.
Specifically, the present embodiment can use OCR (OpticalCharacter Recognition, optical character identification)
Technology carries out text identification, and the answer information of the textual form identified to the answer information of image format.
If 3) answering information is textual form, the present embodiment directly acquires the text shape of information feedback the problem of to output
The answer information of formula.
4) further, the answer information of acquisition is standardized, the corresponding standard of question and answer node is obtained and answers information.
After the aforementioned answer information for obtaining textual form, further answer information can be standardized,
It obtains the corresponding standard of trouble node and answers information.
Under a kind of optional embodiment, question and answer node can also include problem types slot, for storage problem information
Type.The present embodiment can first pass through the various types of the corresponding problem information of destination item in advance, and in corresponding question and answer node
The type of the problem of question and answer node is recorded in problem types slot information.
The type of problem information can there are many, it is common such as whether class problem, description class problem, selection class problem etc..
In the present embodiment, if determining whether the type of the answer information correspondence problem information obtained is according to problem types slot
Class problem to class certainly or negates then class keywords comprising situation according to the answer information of acquisition, determines that standard answers information
For positive or negative.
Specifically, the present embodiment can count in advance affirms class keywords and negative class keywords, affirms class keywords such as:
It is, has ...;Negate class keywords such as: be not that nothing does not have ....
The answer information obtained by matching includes situation to two class keywords, is closed if answering information matches to class certainly
Keyword, it is determined that it is affirmative that standard, which answers information,;If answering information matches to negative class keywords, it is determined that standard answers information
For negative.
Further, if determining that the type of the answer information correspondence problem information obtained is description class according to problem types slot
Problem, then the answer information that will acquire answer information as standard.
Specifically, for description class problem, the answer information that can directly will acquire answers information as standard.
Under another optional embodiment, question and answer node can also include candidate answers slot, believe for storing with problem
Cease matched candidate answers information.Specifically, for certain problem informations, candidate answers information be it is fixed, as problem believe
Breath are as follows: " duration or discontinuity pectoralgia ", corresponding candidate information of answering may include: " duration " and " discontinuity ".
On this basis, if determining that the type of the answer information correspondence problem information obtained is selection according to problem types slot
Class problem can then calculate the similarity of each candidate answers information stored in the answer information and candidate answers slot of acquisition.
Further, according to the size of similarity, determine that standard answers information from candidate answers information.
Specifically, the maximum candidate answers information of similarity can be chosen, answers information as standard, or can be with
The highest topN candidate answers information of similarity is chosen, answers information as standard.
For answer information and candidate answers information similarity calculation process, may include steps of:
A, answer information and candidate answers information are segmented respectively.
When participle, participle model may be constructed.Specifically, answer information training data corresponding to destination item divides
Word mark, and based on annotation results training participle model.Using the participle model after training, answer information and candidate answers are believed
Breath carries out word segmentation processing respectively.
B, stop words removal is carried out to the result after answer information and candidate answers information participle, letter is answered after being handled
Candidate answers information after breath and processing.
C, to candidate answers information computing semantic similarity after answering information after processing and handling.
Specifically, the term vector for each participle that information includes is answered after being handled in term vector model, and every
After one processing in candidate answers information each participle term vector.Further, each point for including according to information is answered after processing
The term vector of word, the term vector for each participle for including with candidate answers information after processing calculate vector distance, as the two
Similarity.
The semantic similarity after information and each processing between candidate answers information is answered after being handled in this step.
In another embodiment of the application, by taking destination item is case acquisition as an example, to step S100, obtain into
The process of the corresponding question and answer node set of destination item of row data acquisition is illustrated, which may include:
S1, the corresponding department's disease of project is acquired according to case, obtain symptom terms relevant to department's disease.
Case acquisition project can correspond to multiple departments, such as internal medicine, surgery.The disease of each department is can in advance really
Fixed, therefore, the corresponding department's disease of project can be acquired in this step according to case, obtains the relevant symptom art of department's disease
Language.
Specifically, it can be obtained from medicine resource data by data digging method and department's disease related symptom term
Set.Medical resource data include relevant medical information on medicine pertinent texts and other networks.Optional mode such as, is pressed
Related disease is obtained from medical text books according to department's title.It further, will be with disease associated description from medicine resource data
Contents extraction comes out.Further, symptom terms label is carried out by description content of the sequence labelling method to extraction, obtains disease
Shape term set.
Wherein, symptom terms are such as: headache, fever, abdominal pain.
Optionally, for the symptom terms set of acquisition, it can therefrom be obtained using frequent set algorithm and be gone out with department disease
The existing highest topM symptom term of frequency.
S2, career in medicine knowledge, which are answered, collects question and answer data relevant to the symptom terms in resource, and is organized into problem information
With answer information.
S3, by the problem information node after arrangement, and according to preset interrogation process by the problem information after node
Form question and answer node set.
Specifically, each problem information can correspond to a question and answer node, can form question and answer node according to interrogation process
Question and answer node set.
Optionally, the present embodiment can in question and answer node offering question class type groove, and by the corresponding type of problem information
It inserts in the problem types slot.
Further alternative, the present embodiment can also be arranged candidate answers slot in question and answer node, and by problem information pair
In the answer information filling candidate answers slot answered.
It is further optional, due to interrogation process be it is determining, collating sequence between question and answer node can also be with
The current question and answer for determining, therefore next question and answer node slot can also being set in question and answer node, and will be determined according to interrogation process
The index of next question and answer node of node is inserted in next question and answer node slot, can be come according to next question and answer node slot so as to subsequent
Determine next question and answer node.
The question and answer node set that the present embodiment generates, can be and store according to tabular form, can also be according to tree-shaped knot
Structure storage, storage form are not specifically limited.
In another embodiment of the application, question and answer are chosen to above-mentioned steps S110, from the question and answer node set
The process of node is illustrated.
The embodiment of the present application discloses several different modes that question and answer node is chosen from question and answer node set, next
Every kind of embodiment is introduced respectively:
The first:
For the corresponding each question and answer node of destination item, its inquiry sequence can be preset.And then the present embodiment can
To choose question and answer node from question and answer node set according to preset inquiry sequence.
Specifically, preset inquiry sequence can be embodied by diversified forms, such as:
1) collating sequence of each question and answer node is kept and inquiry sequence consensus in question and answer node set.It, can be by based on this
According to the collating sequence of question and answer node each in question and answer node set, question and answer node is from first to last chosen.
2) as the aforementioned generating process to question and answer node set is introduced, question and answer node may include next question and answer node
Slot, the index of next question and answer node for storing the question and answer node determined according to the inquiry sequence.Show for example, according to
Inquiry sequence, question and answer node sequencing are as follows: A-B-C-D.Then question and answer can be inserted in next question and answer node slot of question and answer node A
The index of node B.Similarly, for question and answer node B, C, D.
Based on this setup, in the corresponding answer information of question and answer node currently chosen, determination needs to choose
When next question and answer node, the next question and answer section for the next question and answer node slot storage for being included by the question and answer node currently chosen
The corresponding question and answer node of index of point, as next question and answer node.
Second:
For certain form of destination item, corresponding question and answer node possibly can not predefine inquiry sequence.It needs
According to the question and answer node having stepped through, to determine next question and answer node.Based on such destination item, one is present embodiments provided
The scheme that kind is predicted by deep neural network model is as follows in detail:
1) for the problem that each question and answer node chosen in the question and answer node set, believed according to the question and answer node
Breath and answer information, determine the node diagnostic of the question and answer node.
Specifically, question and answer node that definition is currently chosen and that current time had been chosen before is to have chosen question and answer
Node, then for the problem that each has chosen question and answer node, according to having chosen question and answer node information and having answered information, determining should
The node diagnostic of question and answer node is chosen.
Optionally, nodes encoding model can be used to determine the node diagnostic of question and answer node.
Specifically, the project for defining application target project data collected is third party's project.It is disease with destination item
For example acquisition project, disease type can be determined based on the case data of acquisition, then medical diagnosis on disease can be used as third party
Project.For another example, destination item is hearing content acquisition project, can carry out measurement of penalty judgement based on the hearing content of acquisition, then
Measurement of penalty judgement can be used as third party's project.
Based on this, nodes encoding model be can be, and using the question and answer information of destination item and answer information as input data,
And can to input data carry out feature extraction, and according to extraction feature prediction third party's project project result mould
Type.Nodes encoding model can be using the model of two-way length neural network form in short-term, or using the model of other forms.
Based on nodes encoding model, determine that the process of the node diagnostic of question and answer node may include:
A. using information the problem of the question and answer node and answer information as input data, preset nodes encoding mould is inputted
Type;
B. the feature that the nodes encoding model extracts the input data, the node as the question and answer node are obtained
Feature.
In addition to this it is possible to determine the node diagnostic of question and answer node using other modes.As determined question and answer node
Problem information and the corresponding term vector set of answer information, using term vector set as node diagnostic of question and answer node etc..
2) according to the sequencing of selection, the node diagnostic group for each question and answer node chosen is combined into node diagnostic collection
It closes.
Specifically, the node diagnostic of question and answer node can be feature vector form, then can will choose in this step
The feature vector of each question and answer node merges, and according to the sequencing of selection, merges into eigenvectors matrix.
3) the node diagnostic set is inputted to preset node preference pattern, obtains the next of node preference pattern output
The index of question and answer node.
Specifically, the present embodiment can train node preference pattern in advance, corresponding with the destination item when training
The node diagnostic training dataset that the node diagnostic training data of selection question and answer node is combined into according to selection sequence is combined into trained sample
This, using the index of next question and answer node to be chosen of mark as sample label.Node preference pattern can be unidirectional length and remember in short-term
Recall the model of form or the model of other forms.
Based on the node preference pattern after training, node diagnostic set input model can be obtained under model output
The index of one question and answer node.
The output of node preference pattern can be a vector matrix, in the dimension and question and answer node set of vector matrix
The number of question and answer node is identical, and every dimension vector corresponds to unique question and answer node in question and answer node set.Node can be selected
In the vector matrix for selecting model output, the vector that the question and answer node chosen corresponds to dimension is deleted, and in remaining dimension vector
Determine the maximum dimension of vector value, the index by the index of the corresponding question and answer node of the dimension as next question and answer node.
Method provided in this embodiment based on the lower question and answer node of model prediction, it is contemplated that each question and answer node chosen
Node diagnostic, and combine the node preference pattern trained according to training data, be capable of the lower question and answer node of Accurate Prediction
Index.
The third:
Elder generation for certain form of destination item, in corresponding question and answer node set between the question and answer node of possible part
Afterwards inquiry sequence can predefine, and between other parts question and answer node successive inquiry sequence be can not be predetermined.Base
In this, above two implementation can be combined, be can specifically include:
S1, in the corresponding answer information of question and answer node currently chosen, determine when needing to choose next question and answer node,
Judge the rope that next question and answer node whether is stored in next question and answer node slot that the question and answer node currently chosen is included
Draw;If so, S2 is executed, if it is not, executing S3.
Specifically, it if the index for being stored with next question and answer node in next question and answer node slot of question and answer node exists, says
Bright to determine next question and answer node according to predetermined inquiry sequence, otherwise, explanation can not determine, can be based on node
Preference pattern is predicted.
Next question and answer node that S2, the next question and answer node slot for being included by the question and answer node currently chosen store
Corresponding question and answer node is indexed, as next question and answer node.
S3, each question and answer node for the problem that having been chosen in the question and answer node set, according to the question and answer node
Information and answer information, determine the node diagnostic of the question and answer node.
The node diagnostic group for each question and answer node chosen is combined into node diagnostic by S4, the sequencing according to selection
Set.
S5, the node diagnostic set is inputted to preset node preference pattern, obtained under the output of node preference pattern
The index of one question and answer node.
Data acquisition device provided by the embodiments of the present application is described below, data acquisition device described below with
Above-described collecting method can correspond to each other reference.
Referring to fig. 2, Fig. 2 is a kind of data acquisition device structural schematic diagram disclosed in the embodiment of the present application.As shown in Fig. 2,
The apparatus may include:
Question and answer node set acquiring unit 11, for obtaining the corresponding question and answer node of destination item of pending data acquisition
Set, the question and answer node set include question and answer node corresponding with the destination item, and the question and answer node is believed comprising problem
Breath;
Question and answer node selection unit 12, for choosing question and answer node from the question and answer node set;
Problem information output unit 13, for exporting the problem of question and answer node chosen is included information;
Information acquisition unit 14 is answered, for obtaining the answer information to information feedback the problem of output, obtains question and answer section
The corresponding answer information of point.
Optionally, the question and answer node selection unit may include:
Sequentially selection unit, for the inquiry sequence according to the corresponding each question and answer node of the preset destination item, from
Question and answer node is chosen in the question and answer node set.
Optionally, the sequentially selection unit may include:
Sequence selection unit in gathering, for the collating sequence according to each question and answer node in the question and answer node set, from
Head to tail chooses question and answer node, the collating sequence of each question and answer node and the inquiry sequence consensus in the question and answer node set.
Optionally, the question and answer node can also include next question and answer node slot, for storing according to the inquiry sequence
The index of next question and answer node of the determining question and answer node.Based on this, the sequentially selection unit may include:
According to index selection unit, in the corresponding answer information of question and answer node currently chosen, determination to need to select
When removing a question and answer node, the next question and answer for the next question and answer node slot storage for being included by the question and answer node currently chosen
The corresponding question and answer node of the index of node, as next question and answer node.
Optionally, the question and answer node selection unit may include:
Node diagnostic determination unit, for being directed to each question and answer node chosen in the question and answer node set, according to
The problem of question and answer node information and answer information, determine the node diagnostic of the question and answer node;
Feature assembled unit, for the sequencing according to selection, by the node diagnostic for each question and answer node chosen
Group is combined into node diagnostic set;
Node preference pattern predicting unit is obtained for the node diagnostic set to be inputted to preset node preference pattern
The index of the next question and answer node exported to node preference pattern;
The node preference pattern is, with the corresponding node diagnostic training number for having chosen question and answer node of the destination item
It is combined into training sample according to the node diagnostic training dataset being combined into according to selection sequence, with next question and answer node to be chosen of mark
Index be sample label training obtain.
Optionally, the question and answer node can also include next question and answer node slot, for storing the rope of next question and answer node
Draw.Based on this, the question and answer node selection unit can also include:
Question and answer node slot judging unit, in the corresponding answer information of question and answer node currently chosen, determining to be needed
When choosing next question and answer node, judge whether deposit in next question and answer node slot that the question and answer node currently chosen is included
Contain the index of next question and answer node;If so, executing question and answer node slot uses unit, determined if it is not, executing the node diagnostic
Unit;
The question and answer node slot uses unit, next question and answer section for being included by the question and answer node currently chosen
The corresponding question and answer node of index of next question and answer node of point slot storage, as next question and answer node.
Optionally, the node diagnostic determination unit may include:
Nodes encoding model prediction unit, for using information the problem of the question and answer node and answer information as input number
According to, preset nodes encoding model is inputted, the nodes encoding model is, it can carry out feature extraction to input data, and according to
According to the feature of extraction, the model of the project result of third party's project is predicted, third party's project is to apply the destination item
The project of data collected;
Nodes encoding aspect of model extraction unit extracts the input data for obtaining the nodes encoding model
Feature, the node diagnostic as the question and answer node.
Optionally, described problem information output unit may include:
First problem information exports subelement, if being textual form for described problem information, exports in the form of text
The problem of question and answer node of selection is included information, or, the problem of question and answer node by selection is included information carries out voice conjunction
At, and information the problem of export the speech form of synthesis;
Second Problem information exports subelement, if being speech form for described problem information, is exported with speech form
The problem of question and answer node of selection is included information turns or, the problem of question and answer node by selection is included information carries out voice
The problem of writing, and exporting the textual form after transcription information.
Optionally, the answer information acquisition unit may include:
Voice answering acquisition of information subelement, for obtaining the answer letter to the speech form of information feedback the problem of output
Breath, and be the answer information of textual form by its transcription;Or,
Image answers acquisition of information subelement, for obtaining the answer letter to the image format of information feedback the problem of output
Breath, and image text identification is carried out to it, identify the answer information of textual form;Or,
Text answers acquisition of information subelement, for obtaining the answer letter to the textual form of information feedback the problem of output
Breath;
Standardization unit obtains the corresponding standard of question and answer node for being standardized to the answer information of acquisition
Answer information.
Optionally, the question and answer node can also include problem types slot, the type for storage problem information.It is based on
This, the standardization unit may include:
First standardization subelement, if for determining that the answer information correspondence problem obtained is believed according to problem types slot
The type of breath be whether class problem, then to class certainly or negate class keywords comprising situation according to the answers information of acquisition, really
The quasi- information of answering of calibration is positive or negative;
Second standardization subelement, if for determining that the answer information correspondence problem obtained is believed according to problem types slot
The type of breath is description class problem, then the answer information that will acquire answers information as standard.
Optionally, the question and answer node can also include candidate answers slot, for storing and the matched candidate of problem information
Answer information.Based on this, the standardization unit can also include:
Third standardization subelement, if for determining that the answer information correspondence problem obtained is believed according to problem types slot
The type of breath is selection class problem, then calculates each candidate answers information stored in the answer information and candidate answers slot of acquisition
Similarity;According to the size of similarity, determine that standard answers information from candidate answers information.
Data acquisition device provided by the embodiments of the present application can be applied to data acquisition equipment, such as PC terminal, cloud platform, clothes
Business device and server cluster etc..Optionally, Fig. 3 shows the hardware block diagram of data acquisition equipment, and referring to Fig. 3, data are adopted
The hardware configuration for collecting equipment may include: at least one processor 1, at least one communication interface 2,3 He of at least one processor
At least one communication bus 4;
In the embodiment of the present application, processor 1, communication interface 2, memory 3, communication bus 4 quantity be at least one,
And processor 1, communication interface 2, memory 3 complete mutual communication by communication bus 4;
Processor 1 may be a central processor CPU or specific integrated circuit ASIC
(Application Specific Integrated Circuit), or be arranged to implement of the invention real
Apply one or more integrated circuits etc. of example;
Memory 3 may include high speed RAM memory, it is also possible to further include nonvolatile memory (non-volatile
Memory) etc., a for example, at least magnetic disk storage;
Wherein, memory is stored with program, the program that processor can call memory to store, and described program is used for:
Obtain the corresponding question and answer node set of destination item of pending data acquisition, the question and answer node set include with
The corresponding question and answer node of the destination item, the question and answer node include problem information;
Question and answer node is chosen from the question and answer node set, and letter the problem of the question and answer node that exports selection is included
Breath;
The answer information to output problem information feedback is obtained, the corresponding answer information of question and answer node is obtained.
Optionally, the refinement function of described program and extension function can refer to above description.
The embodiment of the present application also provides a kind of readable storage medium storing program for executing, which can be stored with and hold suitable for processor
Capable program, described program are used for:
Obtain the corresponding question and answer node set of destination item of pending data acquisition, the question and answer node set include with
The corresponding question and answer node of the destination item, the question and answer node include problem information;
Question and answer node is chosen from the question and answer node set, and letter the problem of the question and answer node that exports selection is included
Breath;
The answer information to output problem information feedback is obtained, the corresponding answer information of question and answer node is obtained.
Optionally, the refinement function of described program and extension function can refer to above description.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaning
Covering non-exclusive inclusion, so that the process, method, article or equipment for including a series of elements not only includes that
A little elements, but also including other elements that are not explicitly listed, or further include for this process, method, article or
The intrinsic element of equipment.In the absence of more restrictions, the element limited by sentence "including a ...", is not arranged
Except there is also other identical elements in the process, method, article or apparatus that includes the element.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other
The difference of embodiment, the same or similar parts in each embodiment may refer to each other.
The foregoing description of the disclosed embodiments makes professional and technical personnel in the field can be realized or use the application.
Various modifications to these embodiments will be readily apparent to those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the application.Therefore, the application
It is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein phase one
The widest scope of cause.
Claims (15)
1. a kind of collecting method characterized by comprising
Obtain the corresponding question and answer node set of destination item of pending data acquisition, the question and answer node set include with it is described
The corresponding question and answer node of destination item, the question and answer node include problem information;
Question and answer node is chosen from the question and answer node set, and information the problem of the question and answer node that exports selection is included;
The answer information to output problem information feedback is obtained, the corresponding answer information of question and answer node is obtained.
2. the method according to claim 1, wherein described choose question and answer section from the question and answer node set
Point, comprising:
According to the inquiry sequence of the corresponding each question and answer node of the preset destination item, chosen from the question and answer node set
Question and answer node.
3. the method according to claim 1, wherein described choose question and answer section from the question and answer node set
Point, comprising:
For the problem that each question and answer node chosen in the question and answer node set, according to the question and answer node information and return
Information is answered, determines the node diagnostic of the question and answer node;
According to the sequencing of selection, the node diagnostic group for each question and answer node chosen is combined into node diagnostic set;
The node diagnostic set is inputted to preset node preference pattern, obtains next question and answer section of node preference pattern output
The index of point;
The node preference pattern is, with the corresponding node diagnostic training data for having chosen question and answer node of the destination item according to
The node diagnostic training dataset that selection sequence is combined into is combined into training sample, with the rope of next question and answer node to be chosen of mark
Sample label training is cited as to obtain.
4. according to the method described in claim 3, using it is characterized in that, the question and answer node also includes next question and answer node slot
In the index for storing next question and answer node;
It is described that question and answer node is chosen from the question and answer node set, further includes:
Institute is judged when determination needs to choose next question and answer node in the question and answer node corresponding answer information currently chosen
The index that next question and answer node whether is stored in next question and answer node slot that the question and answer node currently chosen is included stated;
If so, the next question and answer node for the next question and answer node slot storage for being included by the question and answer node currently chosen
Corresponding question and answer node is indexed, as next question and answer node;
If it is not, each question and answer node for being directed to and having chosen in the question and answer node set is then executed, according to the question and answer section
The problem of putting information and answer information, determine the operation of the node diagnostic of the question and answer node.
5. according to the method described in claim 3, it is characterized in that, described according to information the problem of the question and answer node and answer
Information determines the node diagnostic of the question and answer node, comprising:
Using information the problem of the question and answer node and information is answered as input data, inputs preset nodes encoding model, institute
Stating nodes encoding model is, feature extraction, and the feature according to extraction can be carried out to input data, prediction third party's project
The model of project result, third party's project are the project using destination item data collected;
Obtain the feature that the nodes encoding model extracts the input data, the node diagnostic as the question and answer node.
6. the method according to claim 1, wherein the answer of the problem of obtaining to output information feedback is believed
Breath, obtains the corresponding answer information of question and answer node, comprising:
The answer information of the speech form of information feedback the problem of to output is obtained, and the answer that its transcription is textual form is believed
Breath;Or,
The answer information of the image format of information feedback the problem of to output is obtained, and image text identification is carried out to it, is identified
The answer information of textual form out;Or,
Obtain the answer information of the textual form of information feedback the problem of to output;
The answer information of acquisition is standardized, the corresponding standard of question and answer node is obtained and answers information.
7. according to the method described in claim 6, it is characterized in that, the question and answer node also includes problem types slot, for depositing
Store up the type of problem information;
The answer information of described pair of acquisition is standardized, and is obtained the corresponding standard of question and answer node and is answered information, comprising:
If according to problem types slot determine obtain answers information correspondence problem information type be whether class problem, basis obtain
The answer information taken is to class certainly or negates class keywords comprising situation, and it is positive or negative that the standard that determines, which answers information,;
If the type for determining the answer information correspondence problem information obtained according to problem types slot is description class problem, will acquire
Answer information as standard answer information.
8. the method according to the description of claim 7 is characterized in that the question and answer node also includes candidate answers slot, for depositing
Storage and the matched candidate answers information of problem information;
The answer information of described pair of acquisition is standardized, and is obtained the corresponding standard of question and answer node and is answered information, further includes:
If the type for determining the answer information correspondence problem information obtained according to problem types slot is selection class problem, calculating is obtained
The similarity of each candidate answers information stored in the answer information and candidate answers slot taken;
According to the size of similarity, determine that standard answers information from candidate answers information.
9. method according to claim 1-8, which is characterized in that the destination item includes case acquisition item
Mesh inquests content acquisition project, interviews any one or more in data acquisition projects.
10. according to the method described in claim 9, it is characterized in that, the destination item is that case acquires project, the then mesh
The generating process of the corresponding question and answer node set of mark project, comprising:
The corresponding department's disease of project is acquired according to case, obtains symptom terms relevant to department's disease;
Career in medicine knowledge, which is answered, collects question and answer data relevant to the symptom terms in resource, and is organized into problem information and answers and believe
Breath;
The problem information after node is formed into question and answer by the problem information node after arrangement, and according to preset interrogation process
Node set.
11. a kind of data acquisition device characterized by comprising
Question and answer node set acquiring unit, for obtaining the corresponding question and answer node set of destination item of pending data acquisition,
The question and answer node set includes question and answer node corresponding with the destination item, and the question and answer node includes problem information;
Question and answer node selection unit, for choosing question and answer node from the question and answer node set;
Problem information output unit, for exporting the problem of question and answer node chosen is included information;
Information acquisition unit is answered, for obtaining the answer information to information feedback the problem of output, it is corresponding to obtain question and answer node
Answer information.
12. device according to claim 11, which is characterized in that the question and answer node selection unit includes:
Node diagnostic determination unit, for being directed to each question and answer node chosen in the question and answer node set, according to described
The problem of question and answer node information and answer information, determine the node diagnostic of the question and answer node;
Feature assembled unit combines the node diagnostic for each question and answer node chosen for the sequencing according to selection
For node diagnostic set;
Node preference pattern predicting unit is saved for the node diagnostic set to be inputted to preset node preference pattern
The index of next question and answer node of point preference pattern output;
The node preference pattern is, with the corresponding node diagnostic training data for having chosen question and answer node of the destination item according to
The node diagnostic training dataset that selection sequence is combined into is combined into training sample, with the rope of next question and answer node to be chosen of mark
Sample label training is cited as to obtain.
13. device according to claim 11, which is characterized in that the answer information acquisition unit includes:
Voice answering acquisition of information subelement, for obtaining the answer information to the speech form of information feedback the problem of output,
And by its transcription be textual form answer information;Or,
Image answers acquisition of information subelement, for obtaining the answer information to the image format of information feedback the problem of output,
And image text identification is carried out to it, identify the answer information of textual form;Or,
Text answers acquisition of information subelement, for obtaining the answer information to the textual form of information feedback the problem of output;
Standardization unit obtains the corresponding standard of question and answer node and answers for being standardized to the answer information of acquisition
Information.
14. a kind of data acquisition equipment, which is characterized in that including memory and processor;
The memory, for storing program;
The processor realizes such as collecting method of any of claims 1-10 for executing described program
Each step.
15. a kind of readable storage medium storing program for executing, is stored thereon with computer program, which is characterized in that the computer program is processed
When device executes, each step such as collecting method of any of claims 1-10 is realized.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810844009.9A CN109065015B (en) | 2018-07-27 | 2018-07-27 | Data acquisition method, device and equipment and readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810844009.9A CN109065015B (en) | 2018-07-27 | 2018-07-27 | Data acquisition method, device and equipment and readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109065015A true CN109065015A (en) | 2018-12-21 |
CN109065015B CN109065015B (en) | 2021-06-08 |
Family
ID=64836766
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810844009.9A Active CN109065015B (en) | 2018-07-27 | 2018-07-27 | Data acquisition method, device and equipment and readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109065015B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110175333A (en) * | 2019-06-04 | 2019-08-27 | 科大讯飞股份有限公司 | A kind of evidence guides method, apparatus, equipment and storage medium |
CN111444701A (en) * | 2019-01-16 | 2020-07-24 | 阿里巴巴集团控股有限公司 | Method and device for prompting inquiry |
CN111667029A (en) * | 2020-07-09 | 2020-09-15 | 腾讯科技(深圳)有限公司 | Clustering method, device, equipment and storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130317816A1 (en) * | 2009-02-02 | 2013-11-28 | Jerry Lee Potter | Method for recognizing and interpreting patterns in noisy data sequences |
CN107133303A (en) * | 2017-04-28 | 2017-09-05 | 百度在线网络技术(北京)有限公司 | Method and apparatus for output information |
CN107403068A (en) * | 2017-07-31 | 2017-11-28 | 合肥工业大学 | Merge the intelligence auxiliary way of inquisition and system of clinical thinking |
CN107562770A (en) * | 2016-07-01 | 2018-01-09 | 松下知识产权经营株式会社 | Information processing method and recording medium |
CN107688608A (en) * | 2017-07-28 | 2018-02-13 | 合肥美的智能科技有限公司 | Intelligent sound answering method, device, computer equipment and readable storage medium storing program for executing |
-
2018
- 2018-07-27 CN CN201810844009.9A patent/CN109065015B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130317816A1 (en) * | 2009-02-02 | 2013-11-28 | Jerry Lee Potter | Method for recognizing and interpreting patterns in noisy data sequences |
CN107562770A (en) * | 2016-07-01 | 2018-01-09 | 松下知识产权经营株式会社 | Information processing method and recording medium |
CN107133303A (en) * | 2017-04-28 | 2017-09-05 | 百度在线网络技术(北京)有限公司 | Method and apparatus for output information |
CN107688608A (en) * | 2017-07-28 | 2018-02-13 | 合肥美的智能科技有限公司 | Intelligent sound answering method, device, computer equipment and readable storage medium storing program for executing |
CN107403068A (en) * | 2017-07-31 | 2017-11-28 | 合肥工业大学 | Merge the intelligence auxiliary way of inquisition and system of clinical thinking |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111444701A (en) * | 2019-01-16 | 2020-07-24 | 阿里巴巴集团控股有限公司 | Method and device for prompting inquiry |
CN110175333A (en) * | 2019-06-04 | 2019-08-27 | 科大讯飞股份有限公司 | A kind of evidence guides method, apparatus, equipment and storage medium |
CN110175333B (en) * | 2019-06-04 | 2023-09-26 | 科大讯飞股份有限公司 | Evidence guiding method, device, equipment and storage medium |
CN111667029A (en) * | 2020-07-09 | 2020-09-15 | 腾讯科技(深圳)有限公司 | Clustering method, device, equipment and storage medium |
CN111667029B (en) * | 2020-07-09 | 2023-11-10 | 腾讯科技(深圳)有限公司 | Clustering method, device, equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN109065015B (en) | 2021-06-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113505586A (en) | Seat-assisted question-answering method and system integrating semantic classification and knowledge graph | |
CN109817046A (en) | A kind of study householder method and private tutor's equipment based on private tutor's equipment | |
CN109065015A (en) | A kind of collecting method, device, equipment and readable storage medium storing program for executing | |
CN106980652A (en) | Intelligent answer method and system | |
CN109902285A (en) | Corpus classification method, device, computer equipment and storage medium | |
CN109446376A (en) | A kind of method and system classified by participle to voice | |
CN109492221A (en) | A kind of information replying method and wearable device based on semantic analysis | |
CN112507139B (en) | Knowledge graph-based question and answer method, system, equipment and storage medium | |
CN110019698A (en) | A kind of intelligent Service method and system of medicine question and answer | |
CN108334501B (en) | Electronic document analysis system and method based on machine learning | |
CN109800309A (en) | Classroom Discourse genre classification methods and device | |
CN108804525A (en) | A kind of intelligent Answering method and device | |
CN108810625A (en) | A kind of control method for playing back of multi-medium data, device and terminal | |
CN110752027A (en) | Electronic medical record data pushing method and device, computer equipment and storage medium | |
CN109766442A (en) | A kind of couple of user takes down notes the method and system classified | |
CN110427470A (en) | Question and answer processing method, device and electronic equipment | |
CN109410656A (en) | It is a kind of that bootstrap technique and facility for study are recited based on melody synthesis | |
CN106844512B (en) | Intelligent question and answer method and system | |
CN112732942A (en) | User-oriented multi-turn question-answer legal document entity relationship extraction method | |
CN109033078B (en) | The recognition methods of sentence classification and device, storage medium, processor | |
CN115455939A (en) | Chapter-level event extraction method, device, equipment and storage medium | |
CN110059179A (en) | A kind of song text name entity recognition method based on deep learning | |
CN115982460A (en) | Personalized recommendation method, system and medium for health science popularization information | |
CN111180076B (en) | Medical information extraction method based on multi-layer semantic analysis | |
US20010049670A1 (en) | Invention interview process |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |