CN108764753A - Test method, apparatus, computer equipment and the storage medium of business personnel's ability - Google Patents
Test method, apparatus, computer equipment and the storage medium of business personnel's ability Download PDFInfo
- Publication number
- CN108764753A CN108764753A CN201810575010.6A CN201810575010A CN108764753A CN 108764753 A CN108764753 A CN 108764753A CN 201810575010 A CN201810575010 A CN 201810575010A CN 108764753 A CN108764753 A CN 108764753A
- Authority
- CN
- China
- Prior art keywords
- business personnel
- business
- ability
- text
- attribute
- 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
- 238000010998 test method Methods 0.000 title description 2
- 230000036651 mood Effects 0.000 claims abstract description 210
- 238000000034 method Methods 0.000 claims abstract description 59
- 238000012360 testing method Methods 0.000 claims abstract description 56
- 238000004458 analytical method Methods 0.000 claims abstract description 20
- 238000012549 training Methods 0.000 claims abstract description 20
- 239000013598 vector Substances 0.000 claims description 26
- 230000015654 memory Effects 0.000 claims description 21
- 241001269238 Data Species 0.000 claims description 14
- 230000007935 neutral effect Effects 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000004088 simulation Methods 0.000 abstract description 11
- 238000010586 diagram Methods 0.000 description 14
- 238000013527 convolutional neural network Methods 0.000 description 12
- 238000012545 processing Methods 0.000 description 9
- 239000008280 blood Substances 0.000 description 6
- 210000004369 blood Anatomy 0.000 description 6
- 238000000205 computational method Methods 0.000 description 6
- 230000008451 emotion Effects 0.000 description 6
- 230000000052 comparative effect Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000013507 mapping Methods 0.000 description 4
- 238000013139 quantization Methods 0.000 description 4
- 238000007792 addition Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000006854 communication Effects 0.000 description 2
- 238000012790 confirmation Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000007787 long-term memory Effects 0.000 description 2
- 230000007115 recruitment Effects 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 206010027951 Mood swings Diseases 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06398—Performance of employee with respect to a job function
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/105—Human resources
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/045—Combinations of networks
-
- 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
- G06N3/08—Learning methods
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Educational Administration (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
This application discloses a kind of method, apparatus, computer equipment and the storage medium of test business personnel's ability, wherein method includes：It controls output device and exports preset question text；Obtain the answer text that business personnel replys described problem text；It will be calculated in the answer text input to the preset sentiment analysis model obtained based on LSTM-CNN model trainings, to obtain the business mood data that the business personnel is directed to described problem text；According to the business mood data, the ability rating of business personnel is determined.The application obtains the problem of simulation client proposes and the reply content of business personnel by the chat of simulation client and business personnel, is to carry out judge business personnel's ability by understanding the business procedure of business personnel to judge the ability of business personnel.
Description
Technical field
This application involves to field of computer technology, especially relate to a kind of method, apparatus of test business personnel's ability,
Computer equipment and storage medium.
Background technology
Now for the professional ability assessment of business personnel, obtained after being largely all based on business personnel and customer communication
The feedback of client and the sales achievement of business personnel are assessed, and are all based on the achievement of business personnel in this way to carry out ability
Assessment.
But business personnel is in the communication process with client, by being exchanged with client, that expresses solves the problems, such as
Ability, the mood to client expressed, these can reflect that the professional ability of business personnel is horizontal.
There is presently no the chats of basic service person and client come the method for judging the ability of business personnel.
Invention content
The main purpose of the application is to provide a kind of chatted by simulation client and business personnel to test business personnel's ability
Method, apparatus, computer equipment and storage medium.
In order to achieve the above-mentioned object of the invention, the application proposes a kind of method of test business personnel's ability, including：
It controls output device and exports preset question text；
Obtain the answer text that business personnel replys described problem text；
By in the answer text input to the preset sentiment analysis model obtained based on LSTM-CNN model trainings into
Row calculates, to obtain the business mood data that the business personnel is directed to described problem text；
According to the business mood data, the ability rating of business personnel is determined.
Further, described according to the business mood data, the step of determining the ability rating of business personnel, including：
The problem of obtaining described problem text mood data, and corresponding problem category is determined according to described problem mood data
Property, wherein described problem attribute includes at least negative attributes, neutral attribute and positive attribute；
According to described problem attribute and the business mood data, determine that the business personnel corresponds to described problem attribute
Ability rating.
Further, the problem of the acquisition question text the step of mood data, including：
It converts described problem text to the structureless vector Z of coding by autocoder, and in the base of the vector Z
Increase structural variable C on plinth, flag sequence is generated using LSTM-RNN methods
By discriminator, by the flag sequenceConvert problematic mood data.
Further, described according to described problem attribute and the business mood data, determine that the business personnel corresponds to
The step of ability rating of described problem attribute, including：
Obtain the corresponding multiple business mood datas of multiple described problem texts of same problem attribute；
It, will be for multiple business mood datas of same described problem attribute and multiple institutes using Regularization formula
State the calculating of problem mood data Regularization；
According to result of calculation, the ability rating of the corresponding business personnel of same described problem attribute is obtained.
Further, described according to described problem attribute and the business mood data, determine that the business personnel corresponds to
After the step of ability rating of described problem attribute, including：
Obtain the highest-capacity grade in the corresponding ability rating of multiple question attributes of the business personnel；
According to the corresponding target problem attribute of the highest-capacity grade, the corresponding post of the target problem attribute is generated
Most suitable post information of the information as the business personnel.
Further, after the ability rating of the determining business personnel the step of includes：
According to the ability rating, the performance score of the business personnel is adjusted.
Further, the business personnel includes multiple；Include after the step of ability rating of the determining business personnel：
Obtain the personal information of the highest target service person of ability rating in the corresponding ability rating of multiple business personnels；
Count the quantity for the characteristic information specified in the corresponding personal information of the multiple business personnel.
The application also provides a kind of device of test business personnel's ability, including：
Output module exports preset question text for controlling output device；
Acquisition module, the answer text that described problem text is replied for obtaining business personnel；
Module is obtained, for by the answer text input to the preset emotion obtained based on LSTM-CNN model trainings
It is calculated in analysis model, to obtain the business mood data that the business personnel is directed to described problem text；
Determining module, for according to the business mood data, determining the ability rating of business personnel.
The application also provides a kind of computer equipment, including memory and processor, and the memory is stored with computer
The step of program, the processor realizes any of the above-described the method when executing the computer program.
The application also provides a kind of computer readable storage medium, is stored thereon with computer program, the computer journey
The step of method described in any one of the above embodiments is realized when sequence is executed by processor.
The application test business personnel's ability method, apparatus, computer equipment and storage medium, by simulate client with
The chat of business personnel, obtain the reply content of the problem of simulation client proposes and business personnel is to judge the ability of business personnel
It carries out judging business personnel's ability by understanding the business procedure of business personnel.When judging business personnel's ability, by problem according to asking
Topic mood is classified, and the corresponding processing capacity understood when business personnel faces different problems more fully judges business personnel's energy
Power.
Description of the drawings
Fig. 1 is the flow diagram of the method for test business personnel's ability of one embodiment of the application；
Fig. 2 illustrates for the detailed process of step S4 in the method for the above-mentioned test business personnel ability of one embodiment of the application
Figure；
Fig. 3 is acquisition question text mood data in the method for the above-mentioned test business personnel ability of one embodiment of the application
Idiographic flow schematic diagram；
Fig. 4 is the idiographic flow schematic diagram of the step S42 in the above-mentioned steps S4 of one embodiment of the application；
Fig. 5 is the flow diagram of the method for test business personnel's ability of one embodiment of the application；
Fig. 6 is the flow diagram of the method for test business personnel's ability of one embodiment of the application；
Fig. 7 is the flow diagram of the method for test business personnel's ability of one embodiment of the application；
Fig. 8 is the structural schematic block diagram of the device of test business personnel's ability of one embodiment of the application；
Fig. 9 is the structural schematic block diagram of the determining module of the device of test business personnel's ability of one embodiment of the application；
Figure 10 is the structural representation picture frame of the acquiring unit of the device of test business personnel's ability of one embodiment of the application；
Figure 11 is the structural schematic block diagram of the determination unit of the device of test business personnel's ability of one embodiment of the application；
Figure 12 is the structural schematic block diagram of the device of test business personnel's ability of one embodiment of the application；
Figure 13 is the structural schematic block diagram of the device of test business personnel's ability of one embodiment of the application；
Figure 14 is the structural schematic block diagram of the device of test business personnel's ability of one embodiment of the application；
Figure 15 is the structural schematic block diagram of the computer equipment of one embodiment of the application.
The embodiments will be further described with reference to the accompanying drawings for realization, functional characteristics and the advantage of the application purpose.
Specific implementation mode
It is with reference to the accompanying drawings and embodiments, right in order to make the object, technical solution and advantage of the application be 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.
Referring to Fig.1, the embodiment of the present application provides a kind of method of test business personnel's ability, including step：
S1, control output device export preset question text；
S2, the answer text that business personnel replys described problem text is obtained；
S3, will be in the answer text input to the preset sentiment analysis model obtained based on LSTM-CNN model trainings
It is calculated, to obtain the business mood data for described problem text of the business personnel；
S4, according to the business mood data, determine the ability rating of business personnel.
As in above-mentioned S1 steps, the main body that this method executes can be a robot, and the output device of robot includes
Display screen and loud speaker.Output device shows preset question text, i.e. robot is when starting to test business personnel, by aobvious
Display screen shows that topic is answered to business personnel, or plays topic by loud speaker and answered to business personnel.The question text is pipe
Some pre-set problems of reason person, or the problem of client proposes for the product of understanding company is simulated, it is organized into text
Form, be stored in robot interior, robot interior is provided with storage card, or is stored in server, and robot passes through
Server is accessed to obtain the question text.When needing to test business personnel, calls out these question texts and shown
Show, it includes directly on the display screen of robot by text that can be, can also be after question text is converted into audio file
It is played out by loud speaker.Different problems text representation different problems.
As in above-mentioned S2 steps, business personnel sees question text, by input units such as keyboard, touch screens to question text institute
The problem of expression, is replied, the answer text that business personnel inputs under robot records.Robot receives the confirmation of business personnel
After information, answer text is stored in memory or server.In another specific embodiment, robot receives business personnel's input
Voice signal after, by speech recognition, convert speech information into text, obtain business personnel reply answer text.
As in above-mentioned S3 steps, LSTM (Long Short-Term Memory) is shot and long term memory network, when being a kind of
Between recurrent neural network, be suitable for processing and predicted time sequence in be spaced and postpone relatively long critical event.LSTM is
Solve the effective technology of long sequence Dependence Problem.Sentiment analysis model is one and LSTM-CNN models is trained rear molds by a large amount of
Type is expressed for the text message of input to be shown that a data, the data are business mood datas after series of computation
The data of the attitude emotion of business personnel.Emotion that is passive and being actively two opposites.Data are bigger, indicate more positive；Data
Smaller, it is more passive to indicate.Sentiment analysis model is to judge mood according to the word of answer text, word.In training emotion point
When analysing model, staff is first by the different business mood data of a large amount of text definition, then by a large amount of text and right
The business mood data answered is separately input in LSTM-CNN models so that the LSTM-CNN models form one according to text
The sentiment analysis model of mood analysis is carried out, robot is by the answer text input of business personnel to the sentiment analysis model later
In after, you can the corresponding business mood data of answer text to export the business personnel.
As in above-mentioned S4 steps, if business mood data is bigger, indicated that the mood of business personnel is more positive, corresponding ability
Higher grade.Business mood data is directly proportional to ability rating.Pass through the mapping of preset business mood data and the grade of service
Rule determines the ability rating of business personnel.
It is further, above-mentioned according to the business mood data with reference to Fig. 2, the step of determining the ability rating of business personnel,
Including：
S41, the problem of described problem text mood data is obtained, and corresponding ask is determined according to described problem mood data
Attribute is inscribed, wherein described problem attribute includes at least negative attributes, neutral attribute and positive attribute；
S42, according to described problem attribute and the business mood data, determine that the business personnel corresponds to described problem category
The ability rating of property.
In the present embodiment, question text is the problem of simulation client proposes, and with certain mood, the feelings of problem
Thread is quantified by problem mood data.According to quantization as a result, each problem is carried out to be respectively defined as different types, often
Corresponding a type is an attribute.Problem mood data corresponding to question text, corresponding attribute of problem mood data etc. this
A little data are that robot is previously stored in RAM card.Robot calls the data in the RAM card, obtains question text
The problem of mood data.Some clients are easier excited, it is desirable that relatively high, the corresponding mood spoken can carry
The problem of passive mood, the question text, mood data was relatively low, and corresponding attribute is negative attributes, further, passive
Attribute includes general negative attributes and unusual negative attributes；Some client's personality are amiable, and mood swing is little, corresponding feelings of speaking
The problem of thread can be relatively flat, question text mood data is medium, and corresponding attribute is neutral attribute；Some clients are relatively more long-pending
Extremely vivaciously, the problem of corresponding mood of speaking, which can compare, carries actively positive mood, question text mood data is relatively high,
Corresponding attribute is positive attribute, and further, positive attribute includes general positive attribute and very positive attribute.Determine business
The ability rating of member, referring to the same business personnel has different ability ratings in face of different attribute questions.Such as following table
It is the corresponding question attributes of the problems in specific embodiment mood data：
In test, robot will at least show one to business personnel respectively the problem of each question attributes, then receive
The reply text that business personnel replys the problem of all problem attributes.Then each business mood number for replying text is calculated
According to calculating belongs to the average value of the corresponding business mood data for replying text of same problem attribute, then by each question attributes
Corresponding mean value calculation summation respectively obtains the corresponding final service mood data of each question attributes of business personnel.Further according to
The corresponding grade of final service mood data carries out the ability rating of evaluation business personnel to business personnel.Each ability rating corresponds to not
Same final service mood data.Following table is the result of a business personnel of robot pair test.
| Question text | Problem mood data | Question attributes | Business mood data |
| Question text 1 | -0.8 | It is very passive | -0.3 |
| Question text 2 | -0.4 | It is general passive | 0.4 |
| Question text 3 | 0.3 | It is neutral | 0.8 |
| Question text 4 | 0.9 | Very actively | 0.9 |
| Question text 5 | 0.4 | It is general positive | 0.8 |
| Question text 6 | 0 | It is neutral | 0 |
| Question text 7 | -0.4 | It is general passive | 0.4 |
In above-described embodiment, according to different problems attribute, identical business mood data may correspond to different abilities
Grade.In one embodiment, as shown above, the corresponding business mood number of 5 question attributes can be calculated separately out
According to average value.For example, question attributes are there are one very passive corresponding business mood datas, and be -0.3, then the problem category
The corresponding business mood data of property is -0.3.Question attributes are there are two general passive corresponding business mood datas, are 0.4 Hes
0.4, then the corresponding business mood data of the question attributes is 0.4.Computational methods according to this show that five question attributes correspond to respectively
Business mood data, be -0.3,0.4,0.4,0.8,0.9.Then this five numerical value are the business personnel of correspondence problem attribute respectively
Ability rating.In one embodiment, ability rating is divided into basic, normal, high three grades, the corresponding business feelings of capabilities grade
Thread data are -0.5 or less (not including -0.5), and the corresponding business mood data of middle ability rating is -0.5 to 0.5, ability etc.
The corresponding business mood data of grade was 0.5 or more (not including 0.5).The then corresponding energy of five question attributes of the business personnel
Power grade be in, in, it is middle and high, high.
In another specific embodiment, it is final service mood data that this five number additions, which are obtained 2.2, i.e., 2.2,.Root again
It is which specific grade of service belonged to according to 2.2, you can to judge which rank business personnel belongs to.Final service mood data
There is a set of preset mapping ruler with the grade of service.
Above-mentioned is a computational methods for calculating final service mood data, can also be calculated most by other computational methods
Whole business mood data.
With reference to Fig. 3, further, the problem of above-mentioned acquisition question text the step of mood data, including：
S411, described problem text is converted by the structureless vector Z of coding by autocoder, and in the vector
Increase structural variable C on the basis of Z, flag sequence is generated using LSTM-RNN methods
S412, by discriminator, by the flag sequenceConvert problematic mood data.
In the present embodiment, the problem of being extracted in the problem of question text is the problem of often proposition from client collection text
Sheet, i.e. text sentence in figure below.By question text by coder transitions at vector, i.e., by text vector.Vectorization
Process can use one-hot Representation models.One-hot Representation be exactly with it is one very long to
It measures to indicate that a word, vector length are the size N of dictionary, there are one each vectors, and dimension is 1, remaining dimension all 0,
Indicate the word in the position of dictionary for 1 position.This One-hot Representation are stored using sparse mode, to
The process of quantization is very succinct.By question text by encoder vectorization after, obtain vector Z, then increase on the basis of Z
Add structural variable C, the purpose for increasing C is to make vector Z can be with to keep the structure of vector Z consistent with subsequent LSTM models
It is input to inside LSTM-RNN models and goes.By obtaining flag sequence after LSTM-RNN mode inputsThen willIt is input to
Discriminator, discriminator identifyMood, obtain the mood data of question text.Wherein, the training process of discriminator is to adopt
With the sentence sample training X with labelL={ (XL, CL), obtain the parameter θ of discriminatorD：
In above-mentioned formula, D represents the sample space of training.After largely trained sample is trained by discriminator, obtain
Generation parameter of the discriminator for the mood data of question text.The problem of obtaining question text when mood data, by problem
Text input to discriminator, the problem of which can be obtained by the formula obtained after training mood data.
It is further, above-mentioned according to described problem attribute and the business mood data with reference to Fig. 4, determine the industry
Business person corresponds to the step of ability rating of described problem attribute, including：
S421, the corresponding multiple business mood datas of multiple described problem texts for obtaining same problem attribute；
S422, using Regularization formula, by for multiple business mood datas of same described problem attribute and
Multiple described problem mood data Regularizations calculate；
S423, according to result of calculation, obtain the ability rating of the corresponding business personnel of same described problem attribute.
In the present embodiment, after business personnel answers multiple problems, mood data can be according to problem number the problem of multiple problems
According to classifying.By the method for cluster, by problem mood data corresponding to multiple question texts of same problem attribute and with
One-to-one business mood data arrange to together, the corresponding business mood data of multiple same problem attributes is utilized into rule
Integralization formula calculates.For example, by exchanging between robot and business personnel, the feelings of business personnel-robot problems text are obtained
Thread tables of data is as follows：
| Problem mood data/Mm | -0.8 | -0.4 | 0.3 | 0.9 | 0.4 | 0 | -0.4 |
| Business mood data/Ym | -0.3 | 0.4 | 0.8 | 0.9 | 0.8 | 0 | 0.4 |
Then Regularization formula is utilized,
Wherein i ∈ j judgment modes：Round function tables, which round up, to be shown, | Ymj|≤1
Regularization obtain following five dimensions it is regular after problem mood data and business mood data：
In this way, business personnel is calculated copes with the business feelings that the corresponding machine mood data of different question texts obtains respectively
Thread data judge the ability rating residing for business personnel respectively to different problems attribute.By after regular by problem mood data
It is classified, it is corresponding that the corresponding business mood data of sorted problem mood data is subjected to regular calculating.
It is further, above-mentioned according to described problem attribute and the business mood data with reference to Fig. 5, determine the industry
After business person corresponds to the step of ability rating of described problem attribute, including：
S5, highest-capacity grade in the corresponding ability rating of multiple question attributes of the business personnel is obtained；
S6, according to the corresponding target problem attribute of the highest-capacity grade, it is corresponding to generate the target problem attribute
Most suitable post information of the post information as the business personnel.
As described in above-mentioned steps S5, text the problem of all problems attribute is sent to business personnel by robot, and then is obtained
To the corresponding ability rating of all problems attribute, there are one ability ratings for each question attributes correspondence.By multiple ability ratings into
Row compares, and obtains highest-capacity grade.Highest-capacity grade can be multiple.
As described in above-mentioned steps S6, target problem attribute refers to question attributes corresponding with highest-capacity grade.Each ask
That inscribe attribute representative is a type of client, and the corresponding ability rating of question attributes is higher, illustrates that business personnel is more good at and is somebody's turn to do
The corresponding people's exchange of question attributes.Different question attributes from client distinguish by text the problem of different Service Periods shows
It is corresponding.Therefore, by business personnel between the corresponding ability rating of different question attributes, the work that business personnel is suitble to can be generated
Post.For example, the problem of highest-capacity grade of first business personnel is corresponding negative attributes attribute, then generate first business personnel and most close
Suitable post information is that relevant post is complained in processing.The problem of highest-capacity grade of second business personnel is corresponding neutral attribute belongs to
Property, then it is that relevant post is seeked advice from foreground to generate the most suitable post information of second business personnel.
With reference to Fig. 6, further, after the ability rating of determining business personnel the step of, includes：
S7, according to the ability rating, adjust the performance score of the business personnel.
As described in above-mentioned steps S7, performance is that tissue is desired as a result, being to be organized as realizing from the point of view of management
Its target and be presented in effective output in different level, the professional ability of one people of achievement performance, the mood number of business personnel
According to higher, illustrate that the enthusiasm of its work is higher, corresponding performance score is also higher.When ability rating is higher than certain grade
When, improve the performance score of business personnel；When ability rating is less than certain grade, the performance score of business personnel is reduced.
With reference to Fig. 7, further, above-mentioned business personnel includes multiple；Step after the ability rating of above-mentioned determining business personnel
Suddenly include：
S8, the personal letter for obtaining the highest target service person of ability rating in the corresponding ability rating of multiple business personnels
Breath, the personal information includes multiple characteristic informations；
The quantity for the characteristic information specified in S9, the corresponding personal information of the multiple business personnel of statistics.
As described in above-mentioned steps S8, robot obtains the personal information that multiple ability ratings are five-star business personnel.It is a
People's information includes multiple features, such as sex character, age characteristics, blood group characteristic, highest educational background feature, native place feature etc..Industry
Business person is before carrying out aptitude tests, and the account information for needing input personal, account information includes just the personal information of business personnel,
Robot calls the personal information in the account information of business personnel.In another specific embodiment, robot is before the test begins
The dialog box of personal information is filled in pop-up after or, then receives the personal information of business personnel's input；After terminating to test, if
Business personnel's ability rating is advanced, then obtains the personal information of business personnel input；If after terminating test, ability rating is not
It is advanced, then the personal information of business personnel input is not obtained.
Include multiple features as described in above-mentioned steps S9, in personal information, each feature includes a characteristic information.Such as
Sex character includes man and the two characteristic informations of female；Age characteristics includes after 70, this after 80s, after 90s, other four spies
Reference ceases；Blood group characteristic includes O, A, B, AB this four characteristic informations.The quantity for counting specific characteristic information, to find out ability
Grade is five-star common trait, facilitates staff in later stage recruiter, the high people of comparative example preferentially recruits.
Further, it calculates after the quantity of specific characteristic information divided by the number of five-star business personnel, obtains each spy
The ratio of reference breath.In one embodiment, 10,000 business personnels are shared and participate in the test, are exported the result is that there are 2,000 etc.
Grade is the business personnel of highest level.Then the personal information of this 2,000 business personnels, one following table of final output are obtained：
After robot counts the ratio, convenient for staff in later stage recruiter, the high people of comparative example preferentially records
It takes.Preferentially enroll the male job candidates of after 90s, O-shaped blood.
Further, equally can also acquisition capability grade lowermost level the corresponding personal information of business personnel, it is each in statistics
The ratio of characteristic information.Convenient for staff recruitment when, the preferentially superseded higher application of characteristic information ratio in this scenario
Personnel.
In conclusion the method for test business personnel's ability of the application is obtained by simulating the chat of client and business personnel
The problem of client proposes and the reply content of business personnel are simulated, is by understanding business personnel to judge the ability of business personnel
Business procedure come carry out judge business personnel's ability.When judging business personnel's ability, problem is classified according to problem mood, it is right
That answers understands processing capacity when business personnel faces different problems, more fully judges business personnel's ability.
With reference to Fig. 8, a kind of device of test business personnel's ability is also provided in the embodiment of the present application, including：
Output module 1 exports preset question text for controlling output device；
First acquisition module 2, the answer text that described problem text is replied for obtaining business personnel；
Module 3 is obtained, for by the answer text input to the preset feelings obtained based on LSTM-CNN model trainings
It is calculated in sense analysis model, to obtain the business mood data that the business personnel is directed to described problem text；
Determining module 4, for according to the business mood data, determining the ability rating of business personnel.
Executive agent in the present embodiment can be a robot.The output device of the output module 1 of robot includes
Display screen and loud speaker.Output device shows preset question text, i.e. robot exports mould when starting to test business personnel
Block 1 shows that topic is answered to business personnel or output module 1 plays topic by loud speaker and returned to business personnel by display screen
It answers.The question text is some pre-set problems of administrator, or simulation client proposes for the product of understanding company
The problem of, it is organized into the form of text, is stored in robot interior, robot interior is provided with storage card, or is stored in
Server, robot obtain the question text by accessing server.When needing to test business personnel, this is called out
A little question texts are shown, can be output module 1 include directly on the display screen of robot by text, can also be defeated
Go out module 1 to be converted into playing out by loud speaker after audio file by question text.Different problems text representation is different
Problem.
Business personnel sees question text, is carried out to the problem expressed by question text by input units such as keyboard, touch screens
It replys, the answer text that business personnel inputs under robot records.After first acquisition module 2 receives the confirmation message of business personnel,
Answer text is stored in memory or server.In another specific embodiment, the first acquisition module 2 receives business personnel's input
Voice signal after, by speech recognition, convert speech information into text, obtain business personnel reply answer text.
LSTM (Long Short-Term Memory) is shot and long term memory network, is a kind of time recurrent neural network,
It is suitable for being spaced and postpone relatively long critical event in processing and predicted time sequence.LSTM is to solve long sequence Dependence Problem
Effective technology.Sentiment analysis model be one by LSTM-CNN models after largely training model, for the text that will input
This information show that a data, the data are business mood datas after series of computation, that is, expresses the attitude emotion of business personnel
Data.Emotion that is passive and being actively two opposites.Data are bigger, indicate more positive；Data are smaller, and it is more passive to indicate.
Sentiment analysis model is to judge mood according to the word of answer text, word.In the training sentiment analysis model, staff
First by the different business mood data of a large amount of text definition, then by a large amount of text and corresponding business mood data point
It is not input in LSTM-CNN models so that the feelings that the LSTM-CNN models form one according to text to carry out mood analysis
Feel analysis model, obtaining module 3 later will be after in the answer text input of business personnel to the sentiment analysis model, you can with output
The corresponding business mood data of answer text of the business personnel.
If business mood data is bigger, indicate that the mood of business personnel is more positive, corresponding ability rating is higher.Business feelings
Thread data are directly proportional to ability rating.By the mapping ruler of preset business mood data and the grade of service, determining module 4 is true
Determine the ability rating of business personnel.
With reference to Fig. 9, further, above-mentioned determining module 4 includes：
Acquiring unit 41, mood data the problem of for obtaining described problem text, and according to described problem mood data
Determining problem attribute, wherein described problem attribute include at least negative attributes, neutral attribute and positive attribute；
Determination unit 42, for according to described problem attribute and the business mood data, determining the business personnel couple
Answer the ability rating of described problem attribute.
In the present embodiment, question text is the problem of simulation client proposes, and with certain mood, the feelings of problem
Thread is quantified by problem mood data.According to quantization as a result, each problem is carried out to be respectively defined as different types, often
Corresponding a type is an attribute.Problem mood data corresponding to question text, corresponding attribute of problem mood data etc. this
A little data are that robot is previously stored in RAM card.Acquiring unit 41 calls the data in the RAM card, obtains problem
The problem of text mood data.Some clients are easier excited, it is desirable that relatively high, the corresponding mood spoken can carry
The problem of some passive moods, the question text, mood data was relatively low, and it is to disappear that acquiring unit 41, which gets corresponding attribute,
Pole attribute, further, negative attributes include general negative attributes and unusual negative attributes；Some client's personality are amiable, mood
The problem of fluctuation is little, and corresponding mood of speaking can be relatively flat, question text mood data is medium, and acquiring unit 41 obtains
It is neutral attribute to corresponding attribute；Some clients are relatively actively more active, and corresponding mood of speaking can compare with actively positive
Mood, mood data is relatively high the problem of the question text, and it is positive attribute that acquiring unit 41, which gets corresponding attribute, into
One step, positive attribute includes general positive attribute and very positive attribute.The ability rating for determining business personnel refers to same
Business personnel has different ability ratings in face of different attribute questions.If following table is the problems in specific embodiment
The corresponding question attributes of mood data：
In test, robot will at least show one to business personnel respectively the problem of each question attributes, then receive
The reply text that business personnel replys the problem of all problem attributes.Then each business mood number for replying text is calculated
According to calculating belongs to the average value of the corresponding business mood data for replying text of same problem attribute, then by each question attributes
Corresponding mean value calculation summation respectively obtains the corresponding final service mood data of each question attributes of business personnel.Further according to
The corresponding grade of final service mood data carries out the ability rating of evaluation business personnel to business personnel.Each ability rating corresponds to not
Same final service mood data.Following table is the result of a business personnel of robot pair test.
| Question text | Problem mood data | Question attributes | Business mood data |
| Question text 1 | -0.8 | It is very passive | -0.3 |
| Question text 2 | -0.4 | It is general passive | 0.4 |
| Question text 3 | 0.3 | It is neutral | 0.8 |
| Question text 4 | 0.9 | Very actively | 0.9 |
| Question text 5 | 0.4 | It is general positive | 0.8 |
| Question text 6 | 0 | It is neutral | 0 |
| Question text 7 | -0.4 | It is general passive | 0.4 |
In above-described embodiment, according to different problems attribute, identical business mood data may correspond to different abilities
Grade.In one embodiment, as shown above, the corresponding business mood number of 5 question attributes can be calculated separately out
According to average value.For example, question attributes are there are one very passive corresponding business mood datas, and be -0.3, then the problem category
The corresponding business mood data of property is -0.3.Question attributes are there are two general passive corresponding business mood datas, are 0.4 Hes
0.4, then the corresponding business mood data of the question attributes is 0.4.Computational methods according to this show that five question attributes correspond to respectively
Business mood data, be -0.3,0.4,0.4,0.8,0.9.Then this five numerical value are the business personnel of correspondence problem attribute respectively
Ability rating.In one embodiment, ability rating is divided into basic, normal, high three grades, the corresponding business feelings of capabilities grade
Thread data are -0.5 or less (not including -0.5), and the corresponding business mood data of middle ability rating is -0.5 to 0.5, ability etc.
The corresponding business mood data of grade was 0.5 or more (not including 0.5).The then corresponding energy of five question attributes of the business personnel
Power grade be in, in, it is middle and high, high.
In another specific embodiment, it is final service mood data that this five number additions, which are obtained 2.2, i.e., 2.2,.Root again
It is which specific grade of service belonged to according to 2.2, determination unit 42 can judge which rank business personnel belongs to.Final industry
Business mood data has a set of preset mapping ruler with the grade of service.
Above-mentioned is a computational methods for calculating final service mood data, can also be calculated most by other computational methods
Whole business mood data.
Referring to Fig.1 0, further, above-mentioned acquiring unit 41 includes：
Sequence subelement 411 converts described problem text to the structureless vector of coding for passing through autocoder
Z, and increase structural variable C on the basis of the vector Z, flag sequence is generated using LSTM-RNN methods
Conversion subunit 412, for passing through discriminator, by the flag sequenceConvert problematic mood data.
In the present embodiment, the problem of being extracted in the problem of question text is the problem of often proposition from client collection text
Sheet, i.e. text sentence in figure below.Sequence subelement 411 by question text by coder transitions at vector, i.e., by text to
Quantization.The process of vectorization can use one-hot Representation models.One-hot Representation are exactly
A word is indicated with a very long vector, vector length is the size N of dictionary, and each vector is only 1 there are one dimension,
Codimension degree all 0 indicates the word in the position of dictionary for 1 position.This One-hot Representation are used
Sparse mode stores, and the process of vectorization is very succinct.Question text is passed through encoder vectorization by sequence subelement 411
Afterwards, obtain vector Z, then increase structural variable C on the basis of Z, the purpose for increasing C be in order to make the structure of vector Z with
Subsequent LSTM models are consistent, so that vector Z is input to inside LSTM-RNN models and go.Pass through LSTM-RNN mode inputs
After obtain flag sequenceThen conversion subunit 412 willIt is input to discriminator, discriminator identifiesMood, obtain
The mood data of question text.Wherein, the training process of discriminator is using the sentence sample training X with labelL={ (XL,
CL), obtain the parameter θ of discriminatorD：
In above-mentioned formula, D represents the sample space of training.After largely trained sample is trained by discriminator, obtain
Generation parameter of the discriminator for the mood data of question text.The problem of obtaining question text when mood data, by problem
Text input to discriminator, the problem of which can be obtained by the formula obtained after training mood data.
Referring to Fig.1 1, further, above-mentioned determination unit 42 includes：
Obtain subelement 421, the corresponding multiple industry of multiple described problem texts for obtaining same problem attribute
Business mood data；
Computation subunit 422 will be for multiple business of same described problem attribute for utilizing Regularization formula
Mood data and multiple described problem mood data Regularizations calculate；
Subelement 423 is obtained, for same according to result of calculation, obtains the corresponding business personnel's of described problem attribute
Ability rating.
In the present embodiment, after business personnel answers multiple problems, mood data can be according to problem number the problem of multiple problems
According to classifying, subelement 421 is obtained by the corresponding multiple business mood numbers of multiple question texts of same problem attribute
According to obtaining simultaneously.By the method for cluster, subelement 421 is obtained by problem corresponding to multiple question texts of same problem attribute
Mood data and therewith one-to-one business mood data, which arrange, to be arrived together, and computation subunit 422 passes through Regularization formula meter
It calculates, the corresponding business mood data Regularization of multiple same problem attributes is calculated.For example, by between robot and business personnel
Exchange, the changeable in mood tables of data for obtaining business personnel-robot problems text is as follows：
| Problem mood data/Mm | -0.8 | -0.4 | 0.3 | 0.9 | 0.4 | 0 | -0.4 |
| Business mood data/Ym | -0.3 | 0.4 | 0.8 | 0.9 | 0.8 | 0 | 0.4 |
Then computation subunit 422 utilizes Regularization formula,
Wherein i ∈ j judgment modes：Round function tables, which round up, to be shown, | Ymj|≤1
Regularization obtain following five dimensions it is regular after problem mood data and business mood data：
| Problem mood data/Mm | -1 | -0.5 | 0 | 0.5 | 1 |
| Business mood data/Ym | -0.500 | 0.275 | 0.5 | 0.937 | 0.944 |
In this way, business personnel is calculated copes with the business feelings that the corresponding machine mood data of different question texts obtains respectively
Thread data obtain subelement 423 respectively to different problems attribute, judge the ability rating residing for business personnel.After regular
Problem mood data is classified, it is corresponding by the corresponding business mood data of sorted problem mood data into professional etiquette
Whole calculating.
Referring to Fig.1 2, further, the device of above-mentioned test business personnel ability further includes：
Second acquisition module 5, in the corresponding ability rating of multiple question attributes for obtaining the business personnel
Highest-capacity grade；
Generation module 6, for according to the corresponding target problem attribute of the highest-capacity grade, generating the target problem
Most suitable post information of the corresponding post information of attribute as the business personnel.
In the present embodiment, target problem attribute refers to question attributes corresponding with highest-capacity grade.By all problems category
The problem of property, text was sent to business personnel, and then obtained the corresponding ability rating of all problems attribute, each question attributes pair
It should be there are one ability rating.Multiple ability ratings are compared, the second acquisition module 5 obtains highest-capacity grade.It can
Power grade can be multiple.That each question attributes represent is a type of client, and the corresponding ability rating of question attributes is got over
Height illustrates that business personnel is more good at people's exchange corresponding with the question attributes.Different question attributes are from client in different business ranks
The problem of section shows text corresponds to respectively.Therefore, by business personnel between the corresponding ability rating of different question attributes,
Generation module 6 can generate the work position that business personnel is suitble to.For example, the highest-capacity grade of first business personnel is corresponding passiveness
The problem of attribute attribute, then it is that relevant post is complained in processing that generation module 6, which generates the most suitable post information of first business personnel,.Second industry
The problem of highest-capacity grade of business person is corresponding neutral attribute attribute, then generation module 6 generate the most suitable hilllock of second business personnel
Position information is that relevant post is seeked advice from foreground.
Referring to Fig.1 3, further, the device of above-mentioned test business personnel ability further includes：
Module 7 is adjusted, for according to the ability rating, adjusting the performance score of the business personnel.
In the present embodiment, performance is that tissue is desired as a result, being to be organized as realizing its target from the point of view of management
And it is presented in effective output in different level, the mood data of the professional ability of one people of achievement performance, business personnel is higher,
Illustrate that the enthusiasm of its work is higher, corresponding performance score is also higher.When ability rating is higher than certain grade, mould is adjusted
Block 7 improves the performance score of business personnel；When ability rating is less than certain grade, adjustment module 7 reduces the performance point of business personnel
Number.
Referring to Fig.1 4, further, above-mentioned business personnel includes multiple；The device of above-mentioned test business personnel ability further includes：
Third acquisition module 8, for obtaining the highest mesh of ability rating in the corresponding ability rating of multiple business personnels
The personal information of business personnel is marked, the personal information includes multiple characteristic informations；
Computing module 9, the quantity for counting the characteristic information specified in the corresponding personal information of the multiple business personnel.
In the present embodiment, third acquisition module 8 obtains the personal information that multiple ability ratings are advanced business personnel.It is personal
Information includes multiple features, such as sex character, age characteristics, blood group characteristic, highest educational background feature, native place feature etc..Business
Member is before carrying out aptitude tests, and the account information for needing input personal, account information includes just the personal information of business personnel, machine
Device people calls the personal information in the account information of business personnel.In another specific embodiment, robot before the test begins or
After pop-up fill in the dialog box of personal information, then receive the personal information of business personnel's input；After terminating to test, if should
Business personnel's ability rating is advanced, then obtains the personal information of business personnel input；If after terminating test, ability rating is not high
Grade does not obtain the personal information of business personnel input then.Include multiple features in personal information, each feature includes a spy
Reference ceases.If sex character includes man and the two characteristic informations of female；Age characteristics includes after 70, is after 80s, after 90s, other
This four characteristic informations；Blood group characteristic includes O, A, B, AB this four characteristic informations.Computing module 9 counts specific characteristic information
Quantity is five-star common trait to find out ability rating, facilitates staff in later stage recruiter, and comparative example is high
People preferentially recruit.
Further, it calculates after the quantity of specific characteristic information divided by the number of five-star business personnel, obtains each spy
The ratio of reference breath.In one embodiment, 10,000 business personnels are shared and participate in the test, are exported the result is that there are 2,000 etc.
Grade is the business personnel of highest level.Then the personal information of this 2,000 business personnels, one following table of final output are obtained：
After robot counts the ratio, convenient for staff in later stage recruiter, the high people of comparative example preferentially records
It takes.Preferentially enroll the male job candidates of after 90s, O-shaped blood.
Further, equally can also acquisition capability grade lowermost level the corresponding personal information of business personnel, it is each in statistics
The ratio of characteristic information.Convenient for staff recruitment when, the preferentially superseded higher application of characteristic information ratio in this scenario
Personnel.
In conclusion the device of test business personnel's ability of the application is obtained by simulating the chat of client and business personnel
The problem of client proposes and the reply content of business personnel are simulated, is by understanding business personnel to judge the ability of business personnel
Business procedure come carry out judge business personnel's ability.When judging business personnel's ability, problem is classified according to problem mood, it is right
That answers understands processing capacity when business personnel faces different problems, more fully judges business personnel's ability.
Referring to Fig.1 5, a kind of computer equipment is also provided in the embodiment of the present application, which can be server,
Its internal structure can be as shown in figure 15.The computer equipment includes processor, memory, the network connected by system bus
Interface and database.Wherein, the processor of the Computer Design is for providing calculating and control ability.The computer equipment is deposited
Reservoir includes non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program
And database.The internal memory provides environment for the operation of operating system and computer program in non-volatile memory medium.It should
The database of computer equipment is used to store the data such as the model of test business personnel's ability.The network interface of the computer equipment is used
It is communicated by network connection in external terminal.To realize a kind of test business personnel when the computer program is executed by processor
The method of ability.
Above-mentioned processor executes the step of method of above-mentioned test business personnel ability：Control output device output is preset to ask
Inscribe text；Obtain the answer text that business personnel replys described problem text；By the answer text input to preset
It is calculated in the sentiment analysis model obtained based on LSTM-CNN model trainings, described problem is directed to obtain the business personnel
The business mood data of text；According to the business mood data, the ability rating of business personnel is determined.
In one embodiment, above-mentioned processor determines the ability rating of business personnel according to the business mood data
Step, including：The problem of obtaining described problem text mood data, and according to problem category corresponding to described problem mood data
Property, wherein described problem attribute includes at least negative attributes, neutral attribute and positive attribute；According to described problem attribute and institute
Business mood data is stated, determines that the business personnel corresponds to the ability rating of described problem attribute.
In one embodiment, above-mentioned processor obtains the step of the problem of question text mood data, including：By certainly
Dynamic encoder converts described problem text to the structureless vector Z of coding, and increases on the basis of the vector Z structural
Variable C generates flag sequence using LSTM-RNN methodsBy discriminator, by the flag sequenceConvert problematic feelings
Thread data.
In one embodiment, above-mentioned processor determines institute according to described problem attribute and the business mood data
The step of business personnel corresponds to the ability rating of described problem attribute is stated, including：Obtain multiple described problems of same problem attribute
The corresponding multiple business mood datas of text；It, will be for multiple institutes of same described problem attribute using Regularization formula
It states business mood data and multiple described problem mood data Regularizations calculates；According to result of calculation, same described ask is obtained
Inscribe the ability rating of the corresponding business personnel of attribute.
In one embodiment, above-mentioned processor determines institute according to described problem attribute and the business mood data
After stating the step of business personnel corresponds to the ability rating of described problem attribute, including：Obtain multiple problem categories of the business personnel
Highest-capacity grade in the corresponding ability rating of property；According to the corresponding target problem attribute of the highest-capacity grade,
Generate most suitable post information of the corresponding post information of the target problem attribute as the business personnel.
In one embodiment, the step of above-mentioned processor determines after the ability rating of business personnel include：According to described
Ability rating adjusts the performance score of the business personnel.
In one embodiment, above-mentioned business personnel includes multiple, and above-mentioned processor determines the step of the ability rating of business personnel
Include after rapid：Obtain the personal letter of the highest target service person of ability rating in the corresponding ability rating of multiple business personnels
Breath, the personal information includes multiple characteristic informations；Count the spy specified in the corresponding personal information of the multiple business personnel
The quantity of reference breath.
In conclusion chat of the computer equipment of the application by simulation client and business personnel, obtains simulation client and carries
The reply content of the problem of going out and business personnel, to judge the ability of business personnel, be by understand business personnel business procedure come
It carries out judging business personnel's ability.When judging business personnel's ability, problem is classified according to problem mood, corresponding understanding industry
Business person faces processing capacity when different problems, more fully judges business personnel's ability.
It will be understood by those skilled in the art that structure shown in Figure 15, only with the relevant part of application scheme
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme.
One embodiment of the application also provides a kind of computer readable storage medium, is stored thereon with computer program, calculates
Machine program realizes a kind of method of test business personnel's ability when being executed by processor, specially：It is default to control output device output
The problem of text；Obtain the answer text that business personnel replys described problem text；By the answer text input in advance
If the sentiment analysis model obtained based on LSTM-CNN model trainings in calculated, to obtain the business personnel for described
The business mood data of question text；According to the business mood data, the ability rating of business personnel is determined.
In one embodiment, above-mentioned processor determines the ability rating of business personnel according to the business mood data
Step, including：The problem of obtaining described problem text mood data, and according to problem category corresponding to described problem mood data
Property, wherein described problem attribute includes at least negative attributes, neutral attribute and positive attribute；According to described problem attribute and institute
Business mood data is stated, determines that the business personnel corresponds to the ability rating of described problem attribute.
In one embodiment, above-mentioned processor obtains the step of the problem of question text mood data, including：By certainly
Dynamic encoder converts described problem text to the structureless vector Z of coding, and increases on the basis of the vector Z structural
Variable C generates flag sequence using LSTM-RNN methodsBy discriminator, by the flag sequenceConvert problematic feelings
Thread data.
In one embodiment, above-mentioned processor determines institute according to described problem attribute and the business mood data
The step of business personnel corresponds to the ability rating of described problem attribute is stated, including：Obtain multiple described problems of same problem attribute
The corresponding multiple business mood datas of text；It, will be for multiple institutes of same described problem attribute using Regularization formula
It states business mood data and multiple described problem mood data Regularizations calculates；According to result of calculation, same described ask is obtained
Inscribe the ability rating of the corresponding business personnel of attribute.
In one embodiment, above-mentioned processor determines institute according to described problem attribute and the business mood data
After stating the step of business personnel corresponds to the ability rating of described problem attribute, including：Obtain multiple problem categories of the business personnel
Highest-capacity grade in the corresponding ability rating of property；According to the corresponding target problem attribute of the highest-capacity grade,
Generate most suitable post information of the corresponding post information of the target problem attribute as the business personnel.
In one embodiment, the step of above-mentioned processor determines after the ability rating of business personnel include：According to described
Ability rating adjusts the performance score of the business personnel.
In one embodiment, above-mentioned business personnel includes multiple, and above-mentioned processor determines the step of the ability rating of business personnel
Include after rapid：Obtain the personal letter of the highest target service person of ability rating in the corresponding ability rating of multiple business personnels
Breath, the personal information includes multiple characteristic informations；Count the spy specified in the corresponding personal information of the multiple business personnel
The quantity of reference breath.
In conclusion chat of the storage medium of the application by simulation client and business personnel, obtains simulation client and proposes
The problem of and business personnel reply content, to judge the ability of business personnel, be by understand business personnel business procedure come into
Row judges business personnel's ability.When judging business personnel's ability, problem is classified according to problem mood, corresponding understanding business
Member faces processing capacity when different problems, more fully judges business personnel's ability.
One of ordinary skill in the art will appreciate that realizing all or part of flow 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 flow of the embodiment of above-mentioned each method.Wherein,
Any reference to memory, storage, database or other media used in provided herein and embodiment,
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 speed are according to rate SDRAM (SSRSDRAM), 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..
It should be noted that herein, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that process, device, article or method including a series of elements include not only those elements, and
And further include other elements that are not explicitly listed, or further include for this process, device, article or method institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including this
There is also other identical elements in the process of element, device, article or method.
The foregoing is merely the preferred embodiments of the application, are not intended to limit the scope of the claims of the application, every utilization
Equivalent structure or equivalent flow shift made by present specification and accompanying drawing content is applied directly or indirectly in other correlations
Technical field, include similarly in the scope of patent protection of the application.
Claims (10)
1. a kind of method of test business personnel's ability, which is characterized in that including：
It controls output device and exports preset question text；
Obtain the answer text that business personnel replys described problem text；
It will be counted in the answer text input to the preset sentiment analysis model obtained based on LSTM-CNN model trainings
It calculates, to obtain the business mood data that the business personnel is directed to described problem text；
According to the business mood data, the ability rating of business personnel is determined.
2. the method for test business personnel's ability as described in claim 1, which is characterized in that described according to the business mood number
According to, the step of determining the ability rating of business personnel, including：
The problem of obtaining described problem text mood data, and corresponding question attributes are determined according to described problem mood data,
Wherein described problem attribute includes at least negative attributes, neutral attribute and positive attribute；
According to described problem attribute and the business mood data, the ability that the business personnel corresponds to described problem attribute is determined
Grade.
3. the method for test business personnel's ability as claimed in claim 2, which is characterized in that the problem of the acquisition question text
The step of mood data, including：
It converts described problem text to the structureless vector Z of coding by autocoder, and on the basis of the vector Z
Increase structural variable C, flag sequence is generated using LSTM-RNN methods
By discriminator, by the flag sequenceConvert problematic mood data.
4. as claimed in claim 2 test business personnel's ability method, which is characterized in that it is described according to described problem attribute with
And the business mood data, determine the step of business personnel corresponds to the ability rating of described problem attribute, including：
Obtain the corresponding multiple business mood datas of multiple described problem texts of same problem attribute；
Using Regularization formula, for multiple business mood datas of same described problem attribute and multiple described will ask
Mood data Regularization is inscribed to calculate；
According to result of calculation, the ability rating of the corresponding business personnel of same described problem attribute is obtained.
5. as claimed in claim 2 test business personnel's ability method, which is characterized in that it is described according to described problem attribute with
And the business mood data, after determining the step of business personnel corresponds to the ability rating of described problem attribute, including：
Obtain the highest-capacity grade in the corresponding ability rating of multiple question attributes of the business personnel；
According to the corresponding target problem attribute of the highest-capacity grade, the corresponding post information of the target problem attribute is generated
Most suitable post information as the business personnel.
6. the method for test business personnel's ability as described in claim 1, which is characterized in that the ability etc. of the determining business personnel
Grade after the step of include：
According to the ability rating, the performance score of the business personnel is adjusted.
7. the method for test business personnel's ability as described in claim 1, which is characterized in that the business personnel includes multiple；Institute
Include after the step of stating the ability rating of determining business personnel：
The personal information of the highest target service person of ability rating in the corresponding ability rating of multiple business personnels is obtained, it is described
Personal information includes multiple characteristic informations；
Count the quantity for the characteristic information specified in the corresponding personal information of the multiple business personnel.
8. a kind of device of test business personnel's ability, which is characterized in that including：
Output module exports preset question text for controlling output device；
Acquisition module, the answer text that described problem text is replied for obtaining business personnel；
Module is obtained, for by the answer text input to the preset sentiment analysis obtained based on LSTM-CNN model trainings
It is calculated in model, to obtain the business mood data that the business personnel is directed to described problem text；
Determining module, for according to the business mood data, determining the ability rating of business personnel.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In when the processor executes the computer program the step of any one of realization claim 1 to 7 the method.
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 claim 1 to 7 is realized when being executed by processor.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810575010.6A CN108764753A (en) | 2018-06-06 | 2018-06-06 | Test method, apparatus, computer equipment and the storage medium of business personnel's ability |
| PCT/CN2018/095025 WO2019232885A1 (en) | 2018-06-06 | 2018-07-09 | Method for testing capabilities of salesman, apparatus, computer device, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810575010.6A CN108764753A (en) | 2018-06-06 | 2018-06-06 | Test method, apparatus, computer equipment and the storage medium of business personnel's ability |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108764753A true CN108764753A (en) | 2018-11-06 |
Family
ID=63999139
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810575010.6A Pending CN108764753A (en) | 2018-06-06 | 2018-06-06 | Test method, apparatus, computer equipment and the storage medium of business personnel's ability |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN108764753A (en) |
| WO (1) | WO2019232885A1 (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109408627A (en) * | 2018-11-15 | 2019-03-01 | 众安信息技术服务有限公司 | A kind of answering method and system merging convolutional neural networks and Recognition with Recurrent Neural Network |
| CN112131369A (en) * | 2020-09-29 | 2020-12-25 | 中国银行股份有限公司 | Service class determination method and device |
| WO2021075950A1 (en) * | 2019-10-14 | 2021-04-22 | Malayan Banking Berhad | Data processing for performance assessment and management |
| CN114781833A (en) * | 2022-04-06 | 2022-07-22 | 平安科技（深圳）有限公司 | Business personnel based capability evaluation method, device, equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105808721A (en) * | 2016-03-07 | 2016-07-27 | 中国科学院声学研究所 | Data mining based customer service content analysis method and system |
| CN107452385A (en) * | 2017-08-16 | 2017-12-08 | 北京世纪好未来教育科技有限公司 | A kind of voice-based data evaluation method and device |
| CN107844889A (en) * | 2017-09-28 | 2018-03-27 | 镇江睿悦读教育科技有限公司 | A kind of system of assessment work ability |
| CN107885726A (en) * | 2017-11-06 | 2018-04-06 | 广州杰赛科技股份有限公司 | Customer service quality evaluating method and device |
| CN107944014A (en) * | 2017-12-11 | 2018-04-20 | 河海大学 | A kind of Chinese text sentiment analysis method based on deep learning |
| US20180114108A1 (en) * | 2016-10-20 | 2018-04-26 | Google Inc. | Answer to question neural networks |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170286867A1 (en) * | 2016-04-05 | 2017-10-05 | Battelle Memorial Institute | Methods to determine likelihood of social media account deletion |
| CN107122351A (en) * | 2017-05-02 | 2017-09-01 | 灯塔财经信息有限公司 | A kind of attitude trend analysis method and system applied to stock news field |
| CN108009756A (en) * | 2017-12-27 | 2018-05-08 | 安徽华久信科技有限公司 | A kind of big data talent ability assessment system and method |
-
2018
- 2018-06-06 CN CN201810575010.6A patent/CN108764753A/en active Pending
- 2018-07-09 WO PCT/CN2018/095025 patent/WO2019232885A1/en active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105808721A (en) * | 2016-03-07 | 2016-07-27 | 中国科学院声学研究所 | Data mining based customer service content analysis method and system |
| US20180114108A1 (en) * | 2016-10-20 | 2018-04-26 | Google Inc. | Answer to question neural networks |
| CN107452385A (en) * | 2017-08-16 | 2017-12-08 | 北京世纪好未来教育科技有限公司 | A kind of voice-based data evaluation method and device |
| CN107844889A (en) * | 2017-09-28 | 2018-03-27 | 镇江睿悦读教育科技有限公司 | A kind of system of assessment work ability |
| CN107885726A (en) * | 2017-11-06 | 2018-04-06 | 广州杰赛科技股份有限公司 | Customer service quality evaluating method and device |
| CN107944014A (en) * | 2017-12-11 | 2018-04-20 | 河海大学 | A kind of Chinese text sentiment analysis method based on deep learning |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109408627A (en) * | 2018-11-15 | 2019-03-01 | 众安信息技术服务有限公司 | A kind of answering method and system merging convolutional neural networks and Recognition with Recurrent Neural Network |
| WO2021075950A1 (en) * | 2019-10-14 | 2021-04-22 | Malayan Banking Berhad | Data processing for performance assessment and management |
| CN112131369A (en) * | 2020-09-29 | 2020-12-25 | 中国银行股份有限公司 | Service class determination method and device |
| CN112131369B (en) * | 2020-09-29 | 2024-02-02 | 中国银行股份有限公司 | Service class determining method and device |
| CN114781833A (en) * | 2022-04-06 | 2022-07-22 | 平安科技（深圳）有限公司 | Business personnel based capability evaluation method, device, equipment and storage medium |
| CN114781833B (en) * | 2022-04-06 | 2024-02-02 | 平安科技（深圳）有限公司 | Capability assessment method, device and equipment based on business personnel and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2019232885A1 (en) | 2019-12-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108764753A (en) | Test method, apparatus, computer equipment and the storage medium of business personnel's ability | |
| CN111160189A (en) | Deep neural network facial expression recognition method based on dynamic target training | |
| CN111985751B (en) | Human-computer chat experience assessment system | |
| CA3109186A1 (en) | Short answer grade prediction | |
| CN112559749B (en) | Intelligent matching method, device and storage medium for online education teachers and students | |
| CN111598153B (en) | Data clustering processing method and device, computer equipment and storage medium | |
| CN110288192A (en) | Quality detecting method, device, equipment and storage medium based on multiple Checking models | |
| CN109086837A (en) | User property classification method, storage medium, device and electronic equipment based on convolutional neural networks | |
| CN112989843B (en) | Intention recognition method, device, computing equipment and storage medium | |
| CN113723774A (en) | Answer scoring method and device, computer equipment and storage medium | |
| CN113570260A (en) | Task allocation method, computer-readable storage medium and electronic device | |
| CN112669936A (en) | Social network depression detection method based on texts and images | |
| CN115859128B (en) | Analysis method and system based on interaction similarity of archive data | |
| US20240005911A1 (en) | Systems and methods to improve trust in conversations with deep learning models | |
| Zhang et al. | Can Large Language Models Assess Personality from Asynchronous Video Interviews? A Comprehensive Evaluation of Validity, Reliability, Fairness, and Rating Patterns | |
| CN114566153A (en) | Voice recognition method and related device, electronic equipment and storage medium | |
| Liang et al. | Intelligent chat robot in digital campus based on deep learning | |
| CN116563067B (en) | Big data-based graduate crowd employment analysis method, system and medium | |
| US20240185832A1 (en) | Systems and methods to automate trust delivery | |
| CN113707185B (en) | Emotion recognition method and device and electronic equipment | |
| CN117934229B (en) | Originality excitation-based talent training guiding method, system, equipment and medium | |
| Chongya | [Retracted] Automatic Scoring of Spoken English Based Principal Component Analysis and Convolutional Neural Network | |
| Indrianti | Face Recognition Application with Deep Neural Networks for Teacher Wellbeing Assessment | |
| Oliveira et al. | Item Response Theory to Evaluate Speech Synthesis: Beyond Synthetic Speech Difficulty. | |
| Škegro et al. | Credit Scoring Analysis: Case Study of Using Weka |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181106 |