CN110275953A - Personality classification method and device - Google Patents
Personality classification method and device Download PDFInfo
- Publication number
- CN110275953A CN110275953A CN201910540702.1A CN201910540702A CN110275953A CN 110275953 A CN110275953 A CN 110275953A CN 201910540702 A CN201910540702 A CN 201910540702A CN 110275953 A CN110275953 A CN 110275953A
- Authority
- CN
- China
- Prior art keywords
- personality
- neural network
- recognition
- recurrent neural
- term vector
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Abstract
Personality classification method provided by the present application and device, obtain test text to be analyzed, and to being pre-processed to test text so that the test text is converted to the term vector that neural network model is capable of handling, and the term vector is inputted Recognition with Recurrent Neural Network.Wherein, after data in the data and personality correlation coefficient charts of presetting network layer output in Recognition with Recurrent Neural Network are spliced, input classification layer obtains the personality classification results that the test text corresponds to author, and personality correlation coefficient charts record has the default degree of correlation between different personal traits.In this way, through the Recognition with Recurrent Neural Network during analyzing test text, in conjunction with the default degree of correlation between different personal traits, so that the personality classification results of prediction are more accurate.
Description
Technical field
This application involves data processing fields, in particular to a kind of personality classification method and device.
Background technique
Personality is the motor tissue of individual inherent psychology physiological system, it determines a people to the unique adaptation of its environment.
There are many personality-formation models in Personality Psychology, in many personality-formation models, the five-factor model personality model of personality school
Because it stablizes, can measure, high confidence level and it is applied widely the advantages that be widely used.Five-factor model personality model includes five people altogether
Lattice speciality, is opening, doing one's duty property, extropism, pleasant property and nervousness respectively, everyone personality type can by this five
A personal traits determines.
Personality classification for text data is to obtain subject by analyzing content of text, such as blog or prose etc.
Content of text corresponds to the personality type of author.Personality classification in generally according to some threshold value by each personality be divided into high personality and
Low two class of personality, wherein the threshold value can be the average mark of the personality, middle number etc..Currently, relatively common way is to be respectively
Each personality establishes two disaggregated models, but this way has ignored the correlation between personal traits, causes classification accuracy low.
Summary of the invention
In order to overcome at least one deficiency in the prior art, the first purpose of the application is to provide a kind of personality classification
Method is applied to data processing equipment, and the data processing equipment is preset with trained Recognition with Recurrent Neural Network, described to train
Recognition with Recurrent Neural Network include people that feature extraction layer, classification layer and record have default degree of correlation between different personal traits
Lattice correlation coefficient charts, which comprises
Obtain the term vector of test text;
The term vector is inputted into the Recognition with Recurrent Neural Network;
By the data in the data and the personality correlation coefficient charts of presetting network layer output in the Recognition with Recurrent Neural Network
Spliced, and inputs the classification layer and obtain the personality type that the test text corresponds to author.
Optionally, the Recognition with Recurrent Neural Network is bidirectional circulating neural network.
Optionally, the method also includes:
For term vector currently entered, the term vector currently entered is obtained by the bidirectional circulating neural network
Feature vector above and following traits vector;
By term vector, feature vector above and the following traits vector currently entered be spliced into new feature to
Amount.
Optionally, the default network layer is maximum pond layer.
Optionally, the default degree of correlation between the different personal traits is related by calculating the Pearson came between personality
It obtains.
Optionally, it is further comprised the steps of: before the term vector for obtaining test text
Word segmentation processing is carried out to the test text, obtains corresponding word segmentation result;
Tool is converted by term vector to handle the word segmentation result, obtains the term vector.
Optionally, the method also includes the training steps to the Recognition with Recurrent Neural Network:
The corresponding term vector of training text is obtained, the term vector of the training text is marked with multiple personal traits labels;
Based on default loss function, the term vector of the training text is inputted into the Recognition with Recurrent Neural Network, by reversed
Propagation algorithm is iterated adjustment to the weight of the Recognition with Recurrent Neural Network, presets until the output valve of the loss function is less than
Threshold value.
The another object of the embodiment of the present application is to provide a kind of personality sorter, applied to data processing equipment, institute
It states data processing equipment and is preset with trained Recognition with Recurrent Neural Network, the trained Recognition with Recurrent Neural Network includes feature extraction
Layer, classification layer and record have the personality correlation coefficient charts of the default degree of correlation between different personal traits, the personality classification
Device includes obtaining module, input module and categorization module;
The term vector for obtaining module and being used to obtain test text;
The input module is used to the term vector inputting the Recognition with Recurrent Neural Network;
The categorization module is used to that the data and the personality phase of network layer output will to be preset in the Recognition with Recurrent Neural Network
The data closed in coefficient table are spliced, and are inputted the classification layer and obtained the personality type that the test text corresponds to author.
Optionally, the personality sorter further includes training module, and the training module is in the following manner to described
Recognition with Recurrent Neural Network is trained:
The corresponding term vector of training text is obtained, the term vector of the training text is marked with multiple personal traits labels;
Based on default loss function, the term vector of the training text is inputted into the Recognition with Recurrent Neural Network, by reversed
Propagation algorithm is iterated adjustment to the weight of the Recognition with Recurrent Neural Network, presets until the output valve of the loss function is less than
Threshold value.
Optionally, the Recognition with Recurrent Neural Network is bidirectional circulating neural network.
In terms of existing technologies, the application has the advantages that
Personality classification method provided by the embodiments of the present application and device obtain test text to be analyzed, and to test
Text is pre-processed so that the test text is converted to the term vector that neural network model is capable of handling, and the term vector is defeated
Enter Recognition with Recurrent Neural Network.It wherein, will be in data and personality correlation coefficient charts that network layer output be preset in Recognition with Recurrent Neural Network
After data are spliced, input classification layer obtains the personality classification results that the test text corresponds to author, the personality related coefficient
Table record has the default degree of correlation between different personal traits.In this way, by the Recognition with Recurrent Neural Network in analysis test text
In the process, in conjunction with the default degree of correlation between different personal traits, so that the personality classification results of prediction are more accurate.
Detailed description of the invention
Technical solution in ord to more clearly illustrate embodiments of the present application, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only some embodiments of the application, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 is the hardware structure diagram of data processing equipment provided by the embodiments of the present application;
Fig. 2 is the step flow chart of personality classification method provided by the embodiments of the present application;
Fig. 3 is the circuit theory schematic diagram of Recognition with Recurrent Neural Network provided by the embodiments of the present application;
Fig. 4 is personality correlation coefficient charts provided by the embodiments of the present application;
Fig. 5 is one of the structural schematic diagram of personality sorter provided by the embodiments of the present application;
Fig. 6 is the second structural representation of personality sorter provided by the embodiments of the present application.
Icon: 100- data processing equipment;130- processor;110- personality sorter;120- memory;501- recurrence
Layer;The pond 502- layer;The full articulamentum of 503-;504-softmax layers;505- personality correlation coefficient charts;1101- obtains module;
1102- input module;1103- categorization module;1104- training module.
Specific embodiment
To keep the purposes, technical schemes and advantages of the embodiment of the present application clearer, below in conjunction with the embodiment of the present application
In attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is
Some embodiments of the present application, instead of all the embodiments.The application being usually described and illustrated herein in the accompanying drawings is implemented
The component of example can be arranged and be designed with a variety of different configurations.
Therefore, the detailed description of the embodiments herein provided in the accompanying drawings is not intended to limit below claimed
Scope of the present application, but be merely representative of the selected embodiment of the application.Based on the embodiment in the application, this field is common
Technical staff's every other embodiment obtained without creative efforts belongs to the model of the application protection
It encloses.
It should also be noted that similar label and letter indicate similar terms in following attached drawing, therefore, once a certain Xiang Yi
It is defined in a attached drawing, does not then need that it is further defined and explained in subsequent attached drawing.
Please refer to Fig. 1, Fig. 1 is the hardware structure diagram of data processing equipment 100 provided by the embodiments of the present application, at the data
Managing equipment 100 includes processor 130, memory 120 and personality sorter 110.
The memory 120 and each element of processor 130 are directly or indirectly electrically connected between each other, to realize data
Transmission or interaction.Electrically connect for example, these elements can be realized between each other by one or more communication bus or signal wire
It connects.The personality sorter 110 includes that at least one can be stored in described deposit in the form of software or firmware (firmware)
In reservoir 120 or the software function that is solidificated in the operating system (operating system, OS) of the data processing equipment 100
It can module.The processor 130 is for executing the executable module stored in the memory 120, such as personality classification
Software function module included by device 110 and computer program etc..
Wherein, the memory 120 may be, but not limited to, random access memory (Random Access
Memory, RAM), read-only memory (Read Only Memory, ROM), programmable read only memory (Programmable
Read-Only Memory, PROM), erasable read-only memory (Erasable Programmable Read-Only
Memory, EPROM), electricallyerasable ROM (EEROM) (Electric Erasable Programmable Read-Only
Memory, EEPROM) etc..Wherein, memory 120 is for storing program, the processor 130 after receiving and executing instruction,
Execute described program.
The processor 130 may be a kind of IC chip, the processing capacity with signal.Above-mentioned processor can
To be general processor, including central processing unit (Central Processing Unit, abbreviation CPU), network processing unit
(Network Processor, abbreviation NP) etc.;Can also be digital signal processor (DSP), specific integrated circuit (ASIC),
Field programmable gate array (FPGA) either other programmable logic device, discrete gate or transistor logic, discrete hard
Part component.It may be implemented or execute disclosed each method, step and the logic diagram in the embodiment of the present application.General processor
It can be microprocessor or the processor be also possible to any conventional processor etc..
The data processing equipment 100 may be, but not limited to, smart phone, PC (personal
Computer, PC), tablet computer, personal digital assistant (personal digital assistant, PDA), mobile Internet access set
Standby (mobile Internet device, MID) etc..
Referring to figure 2., Fig. 2 is a kind of process of personality classification method applied to data processing equipment 100 shown in FIG. 1
Figure, the data processing equipment 100 are preset with trained Recognition with Recurrent Neural Network.Referring to figure 3., Fig. 3 mentions for the embodiment of the present application
The network structure of the Recognition with Recurrent Neural Network of confession.The Recognition with Recurrent Neural Network includes that feature extraction layer, classification layer and record are different
The personality correlation coefficient charts 505 of default degree of correlation between personal traits;Wherein, which includes complete 503 He of articulamentum
Softmax layer 504;This feature extract layer includes recurrence layer 501 and pond layer 502.It below will be to the method includes each steps
Suddenly it is described in detail.
Step S100 obtains the term vector of text to be tested.
Optionally, which can obtain a large amount of test text, the test from local or network
Text can be blog, prose, diary or composition etc..Before the test text is inputted Recognition with Recurrent Neural Network, need pair
The test text is pre-processed so that test text is converted to the term vector that Recognition with Recurrent Neural Network is capable of handling.
For example, in a kind of possible example, the data processing equipment 100 by vocabulary or dictionary to test text into
Row word segmentation processing.It is worth noting that if the test text is the text data of Chinese class, since Chinese is the same different from English
Word is distinguished by space.Therefore before data processing equipment 100 is to test text processing, it is necessary first to test
Text carries out word segmentation processing.The quality of word segmentation processing often will affect the analysis result to the test text.
For example, to " today, weather was fine!" word segmentation processing is carried out, good word segmentation result is " today " " weather " " very good ",
And the word segmentation result of difference is " the present " " everyday " " gas " " very good ".As can be seen that different word segmentation results, entirely different by bringing
The semanteme meaning.
Test text after participle is carried out one-hot coding by the data processing equipment 100, i.e., how many a states just have
How many corresponding bits.For example, to " today, weather was very good!" " very good " the progress one- of word segmentation result " today " " weather "
Hot coding, which is corresponding with 3 states, therefore is corresponding with 3 bits.The coding result of " today " is " 100 ";
The coding result of " weather " is " 010 ";The coding result of " very good " is " 001 ".
The data processing equipment 100 obtains the term vector of the test text by searching for the good term vector table of pre-training.Example
Such as, the term vector of " today " is [0.2,0.3];The term vector of " weather " is [0.4,0.8], the term vector of " very good " be [0.5,
0.9].Wherein, the term vector table of pre-training is that training obtains in external corpus using the skip-gram in word2vec.
The term vector is inputted the Recognition with Recurrent Neural Network by step S200.
Step S300 will preset the data and the personality correlation coefficient charts of network layer output in the Recognition with Recurrent Neural Network
Data in 505 are spliced, and are inputted the classification layer and obtained the personality type that the test text corresponds to author.
Optionally, it is worth noting that, often in text data a certain position semantic information, with text data up and down
Text is related, therefore the Recognition with Recurrent Neural Network in the present embodiment can be bidirectional circulating neural network.The Recognition with Recurrent Neural Network is
Bidirectional circulating neural network (Bidirectional Recurrent Neural Networks, BRNN) can be good at handling
Contextual information in text data.
For example, " my mobile phone is broken, I intend () new cell-phone ", predicts the word that should be inserted in bracket, from including
Information number above, can insert " buying " and " repairing " in the bracket or mobile phone is broken causes to feel blue, and can fill out in bracket
Enter " crying ", " relieving boredom for a walk " and " having a big dinner ".But if considering the subsequent information of bracket, filling in the bracket
A possibility that " buying ", is bigger.
Based on this thought, before the output of bidirectional circulating neural network current time i not only relies in sequence when i-1
The input at quarter also relies on the input at subsequent i+1 moment.For term vector currently entered, which is mentioned
Take the feature vector above and following traits vector of current term vector, wherein feature vector c abovel(wi) can indicate are as follows:
cl(wi)=f (W(l)cl(wi-1)+W(sl)e(wi-1));
Wherein, cl(wi-1) exported for the forward direction of the i-1 moment bidirectional circulating neural network recurrence layer 501, e (wi-1) it is i-
The term vector of the 1 moment bidirectional circulating neural network, W(l)And W(sl)Respectively its corresponding weight.
Following traits vector cr(wi) can indicate are as follows:
cr(wi)=f (W(r)cr(wi+1)+W(sr)e(wi+1));
Wherein, cr(wi+1) be the i+1 moment bidirectional circulating neural network recurrence layer 501 backward output, e (wi+1) it is i+
The term vector of the 1 moment bidirectional circulating neural network, W(r)With W(sr)Respectively its corresponding weight.The data processing equipment 100
Feature vector above and the splicing of following traits vector are obtained into current semantic feature xi:
xi=[cl(wi);e(wi);cr(wi)];
Wherein, e (wi) it is current term vector, implicit semantic is obtained in the following way
WhIt is xiWeight, bhIt is xiBiasing, tanh () is the activation primitive of hidden layer, calculation formula are as follows:
Optionally, which is maximum pond layer, and data processing equipment 100 is by the output feature of recurrence layer 501
By the maximum pond, layer carries out processing and obtains ypool.Calculation are as follows:
By ypoolSplicing, which is carried out, with the data in personality correlation coefficient charts 505 obtains xf, it is then input to full articulamentum 503
In, calculation is as follows:
xf=[ypool;r];
yf=Wfxf+bf;
Wherein, r is the default degree of correlation between the different personal traits recorded in personality correlation coefficient charts 505, different
Default degree of correlation between personal traits is obtained by the Pearson came correlation calculated between personality.WfIt is xfWeight matrix, bf
It is xfBiasing, yfIt is the output of full articulamentum 503.
For example, in a kind of possible example, for indicate opening, doing one's duty property, pleasant property, it is export-oriented and it is neurotic it
Between degree of correlation personality correlation coefficient charts 505 it is as shown in Figure 4.Degree of correlation between personality is by " related coefficient " and " aobvious
Work property " determines." double tails " indicates a kind of measurement criteria in figure, and there are also " single tails " for corresponding measurement criteria.It is this at " double tails "
It in measurement criteria, when significance is greater than 0.01 less than 0.05, is indicated with one " * ", if significance is used less than 0.01
It is indicated in two " * * ".As shown in figure 4, the related coefficient between " opening " and " doing one's duty property " is 0.29, and between the two
Conspicuousness is greater than 0.01 less than 0.05.
Data processing equipment 100 by personality correlation coefficient charts 505 related coefficient and conspicuousness parameter produce in column
Vector, and carry out processing with maximum pond layer and obtain ypoolSpliced, and the full articulamentum 503 inputted in classification layer is somebody's turn to do
Test text corresponds to the classification results of the personality type of author.Wherein, the output of the full articulamentum 503 is connected with softmax layers
504, which is normalized by the data that 504 pairs of softmax layer full articulamentums 503 export, and is obtained
It obtains test text and corresponds to probability and its probability threshold value that each speciality of author belongs to high speciality.
For example, softmax layers of output shares 10 outputs in a kind of possible example, it is divided into 5 personal traits
Personality probability and the corresponding probability threshold value of 5 personal traits.Wherein, each personal traits can be divided into high speciality and
Low speciality, as extropism can be divided into high extropism and low extropism.If the personality probability of a personality is more than or equal to its correspondence
Probability threshold value, then the personal traits be Gao Tezhi;If being less than its corresponding probability threshold value, which is low speciality.
Wherein, the calculation of softmax is as follows:
If softmax exports as { 0.05,0.1,0.16,0.13,0.06,0.11,0.04,0.09,0.14,0.12 } wherein
0.05 is the probability that author has high opening, and 0.1 is high open threshold value, and 0.05 < 0.1 so author does not have high open
Property, that is, author is with low opening;0.16 is the probability that author has high doing one's duty property, 0.16 > 0.13, so, author's tool
There is doing one's duty property of height.
Optionally, the embodiment of the present application also provides a kind of training method to the Recognition with Recurrent Neural Network, the training method
Step includes:
The data processing equipment 100 obtains the corresponding term vector of training text, and the term vector of the training text is marked with more
A personal traits label.Wherein, it before the term vector for obtaining training text, needs to pre-process training text, obtain
The term vector of the training text.Pretreated method includes first carrying out word segmentation processing to the training text, after word segmentation processing
Data carry out one-hot coding, then the term vector table good by searching for pre-training, by the training text of one-hot coding form
Originally it is converted to corresponding term vector.In the embodiment of the present application, the term vector table of the pre-training is using the skip- in word2vec
Gram training in external corpus obtains.
Based on default loss function, the term vector of the training text is inputted into the Recognition with Recurrent Neural Network, by reversed
Propagation algorithm is iterated adjustment to the weight of the Recognition with Recurrent Neural Network, presets until the output valve of the loss function is less than
Threshold value.Wherein, the calculation of the default loss function is as follows:
Wherein,It is training text diRelated personality.For example, if training text diThe personality of corresponding author is height
Doing one's duty property and height are neurotic, then doing one's duty property, nervousness are diRelated personality;So remaining three personalities be it is open,
Extropism and pleasant property are diUncorrelated personality,It isSupplementary set, i.e. diUncorrelated personality.It is training text di
The neuron output,It is training text diThe output probability of related personal traits label,It is training text
diThe output probability of uncorrelated personal traits label,WithBetween gap be the bigger the better,It is training text di
The threshold value of related personal traits label 2k, output probability are the bigger the better higher than threshold value.It is training text dlUncorrelated personality
The threshold value of speciality label 2j.
The application is that embodiment also provides a kind of personality sorter 110.Referring to figure 5., Fig. 5 is the personality sorter
110 structural schematic diagram, is applied to data processing equipment 100, and the data processing equipment 100 is preset with trained circulation mind
Through network, the trained Recognition with Recurrent Neural Network includes that feature extraction layer, classification layer and record have between different personal traits
Default degree of correlation personality correlation coefficient charts 505, the personality sorter 110 include obtain module 1101, input mould
Block 1102 and categorization module 1103.
The term vector for obtaining module 1101 and being used to obtain test text.
In the embodiment of the present application, which is used to execute the step S100 in Fig. 2, about the acquisition module
1101 detailed description can refer to the detailed description of step S100.
The input module 1102 is used to the term vector inputting the Recognition with Recurrent Neural Network.
In the embodiment of the present application, which is used to execute the step S200 in Fig. 2, about the input module
1102 detailed description can refer to the detailed description of step S200.
The categorization module 1103 is used for the data for presetting network layer output in the Recognition with Recurrent Neural Network and the people
Data in lattice correlation coefficient charts 505 are spliced, and are inputted the classification layer and obtained the people that the test text corresponds to author
Lattice type.
In the embodiment of the present application, which is used to execute the step S300 in Fig. 2, about the categorization module
1103 detailed description can refer to the detailed description of step S300.
Fig. 6 is please referred to, which further includes training module 1104, which passes through following
Mode is trained the Recognition with Recurrent Neural Network:
The corresponding term vector of training text is obtained, the term vector of the training text is marked with multiple personal traits labels;
Based on default loss function, the term vector of the training text is inputted into the Recognition with Recurrent Neural Network, by reversed
Propagation algorithm is iterated adjustment to the weight of the Recognition with Recurrent Neural Network, presets until the output valve of the loss function is less than
Threshold value.
The Recognition with Recurrent Neural Network can be bidirectional circulating neural network.
In conclusion personality classification method provided by the embodiments of the present application and device, obtain test text to be analyzed, and
To being pre-processed to test text so that the test text is converted to the term vector that neural network model is capable of handling, and should
Term vector inputs Recognition with Recurrent Neural Network.Wherein, the data and personality phase relation of network layer output will be preset in Recognition with Recurrent Neural Network
After data in number table are spliced, input classification layer obtains the personality classification results that the test text corresponds to author, the personality
Correlation coefficient charts record has the default degree of correlation between different personal traits.In this way, being surveyed by the Recognition with Recurrent Neural Network in analysis
During trying text, in conjunction with the default degree of correlation between different personal traits, so that the personality classification results of prediction are more quasi-
Really.
In embodiment provided herein, it should be understood that disclosed device and method, it can also be by other
Mode realize.The apparatus embodiments described above are merely exemplary, for example, the flow chart and block diagram in attached drawing are shown
According to device, the architectural framework in the cards of method and computer program product, function of multiple embodiments of the application
And operation.In this regard, each box in flowchart or block diagram can represent one of a module, section or code
Point, a part of the module, section or code includes one or more for implementing the specified logical function executable
Instruction.It should also be noted that function marked in the box can also be attached to be different from some implementations as replacement
The sequence marked in figure occurs.For example, two continuous boxes can actually be basically executed in parallel, they sometimes may be used
To execute in the opposite order, this depends on the function involved.It is also noted that each of block diagram and or flow chart
The combination of box in box and block diagram and or flow chart can be based on the defined function of execution or the dedicated of movement
The system of hardware is realized, or can be realized using a combination of dedicated hardware and computer instructions.
In addition, each functional module in each embodiment of the application can integrate one independent portion of formation together
Point, it is also possible to modules individualism, an independent part can also be integrated to form with two or more modules.
It, can be with if the function is realized and when sold or used as an independent product in the form of software function module
It is stored in a computer readable storage medium.Based on this understanding, the technical solution of the application is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be embodied in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be a
People's computer, server or network equipment etc.) execute each embodiment the method for the application all or part of the steps.
And storage medium above-mentioned includes: that USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
The various media that can store program code such as reservoir (RAM, Random Access Memory), magnetic or disk.
It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
The above, the only various embodiments of the application, but the protection scope of the application is not limited thereto, it is any
Those familiar with the art within the technical scope of the present application, can easily think of the change or the replacement, and should all contain
Lid is within the scope of protection of this application.Therefore, the protection scope of the application shall be subject to the protection scope of the claim.
Claims (10)
1. a kind of personality classification method, which is characterized in that be applied to data processing equipment, the data processing equipment is preset with instruction
The Recognition with Recurrent Neural Network perfected, the trained Recognition with Recurrent Neural Network include that feature extraction layer, classification layer and record are different
The personality correlation coefficient charts of default degree of correlation between personal traits, which comprises
Obtain the term vector of test text;
The term vector is inputted into the Recognition with Recurrent Neural Network;
Data in the data and the personality correlation coefficient charts of presetting network layer output in the Recognition with Recurrent Neural Network are carried out
Splicing, and input the classification layer and obtain the personality type that the test text corresponds to author.
2. personality classification method according to claim 1, which is characterized in that the Recognition with Recurrent Neural Network is bidirectional circulating mind
Through network.
3. personality classification method according to claim 2, which is characterized in that the method also includes:
For term vector currently entered, the upper of the term vector currently entered is obtained by the bidirectional circulating neural network
Literary feature vector and following traits vector;
Term vector, feature vector above and the following traits vector currently entered are spliced into new feature vector.
4. personality classification method according to claim 3, which is characterized in that the default network layer is maximum pond layer.
5. personality classification method according to claim 1, which is characterized in that the default phase between the difference personal traits
Pass degree is obtained by the Pearson came correlation calculated between personal traits.
6. personality classification method according to claim 1, which is characterized in that before the term vector for obtaining test text
It further comprises the steps of:
Word segmentation processing is carried out to the test text, obtains corresponding word segmentation result;
Tool is converted by term vector to handle the word segmentation result, obtains the term vector.
7. personality classification method according to claim 1, which is characterized in that the method also includes to the circulation nerve
The training step of network:
The corresponding term vector of training text is obtained, the term vector of the training text is marked with multiple personal traits labels;
Based on default loss function, the term vector of the training text is inputted into the Recognition with Recurrent Neural Network, passes through backpropagation
Algorithm is iterated adjustment to the weight of the Recognition with Recurrent Neural Network, until the output valve of the loss function is less than default threshold
Value.
8. a kind of personality sorter, which is characterized in that be applied to data processing equipment, the data processing equipment is preset with instruction
The Recognition with Recurrent Neural Network perfected, the trained Recognition with Recurrent Neural Network include that feature extraction layer, classification layer and record are different
The personality correlation coefficient charts of default degree of correlation between personal traits, the personality sorter include obtaining module, input
Module and categorization module;
The term vector for obtaining module and being used to obtain test text;
The input module is used to the term vector inputting the Recognition with Recurrent Neural Network;
The categorization module is used to that the data and the personality phase relation of network layer output will to be preset in the Recognition with Recurrent Neural Network
Data in number table are spliced, and are inputted the classification layer and obtained the personality type that the test text corresponds to author.
9. personality sorter according to claim 8, which is characterized in that the personality sorter further includes trained mould
Block, the training module are in the following manner trained the Recognition with Recurrent Neural Network:
The corresponding term vector of training text is obtained, the term vector of the training text is marked with multiple personal traits labels;
Based on default loss function, the term vector of the training text is inputted into the Recognition with Recurrent Neural Network, passes through backpropagation
Algorithm is iterated adjustment to the weight of the Recognition with Recurrent Neural Network, until the output valve of the loss function is less than default threshold
Value.
10. personality sorter according to claim 8, which is characterized in that the Recognition with Recurrent Neural Network is bidirectional circulating
Neural network.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910540702.1A CN110275953B (en) | 2019-06-21 | 2019-06-21 | Personality classification method and apparatus |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910540702.1A CN110275953B (en) | 2019-06-21 | 2019-06-21 | Personality classification method and apparatus |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110275953A true CN110275953A (en) | 2019-09-24 |
CN110275953B CN110275953B (en) | 2021-11-30 |
Family
ID=67961812
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910540702.1A Active CN110275953B (en) | 2019-06-21 | 2019-06-21 | Personality classification method and apparatus |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110275953B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112487184A (en) * | 2020-11-26 | 2021-03-12 | 北京智源人工智能研究院 | User character judging method and device, memory and electronic equipment |
CN113221560A (en) * | 2021-05-31 | 2021-08-06 | 平安科技(深圳)有限公司 | Personality trait and emotion prediction method, personality trait and emotion prediction device, computer device, and medium |
CN113268740A (en) * | 2021-05-27 | 2021-08-17 | 四川大学 | Input constraint completeness detection method of website system |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104615589A (en) * | 2015-02-15 | 2015-05-13 | 百度在线网络技术(北京)有限公司 | Named-entity recognition model training method and named-entity recognition method and device |
CN105701460A (en) * | 2016-01-07 | 2016-06-22 | 王跃明 | Video-based basketball goal detection method and device |
CN105975504A (en) * | 2016-04-28 | 2016-09-28 | 中国科学院计算技术研究所 | Recurrent neural network-based social network message burst detection method and system |
CN107451118A (en) * | 2017-07-21 | 2017-12-08 | 西安电子科技大学 | Sentence-level sensibility classification method based on Weakly supervised deep learning |
CN108108354A (en) * | 2017-06-18 | 2018-06-01 | 北京理工大学 | A kind of microblog users gender prediction's method based on deep learning |
US20180203848A1 (en) * | 2017-01-17 | 2018-07-19 | Xerox Corporation | Author personality trait recognition from short texts with a deep compositional learning approach |
US10169656B2 (en) * | 2016-08-29 | 2019-01-01 | Nec Corporation | Video system using dual stage attention based recurrent neural network for future event prediction |
CN109376784A (en) * | 2018-10-29 | 2019-02-22 | 四川大学 | A kind of personality prediction technique and personality prediction meanss |
CN109597891A (en) * | 2018-11-26 | 2019-04-09 | 重庆邮电大学 | Text emotion analysis method based on two-way length Memory Neural Networks in short-term |
CN109829154A (en) * | 2019-01-16 | 2019-05-31 | 中南民族大学 | Semantic-based personality prediction technique, user equipment, storage medium and device |
-
2019
- 2019-06-21 CN CN201910540702.1A patent/CN110275953B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104615589A (en) * | 2015-02-15 | 2015-05-13 | 百度在线网络技术(北京)有限公司 | Named-entity recognition model training method and named-entity recognition method and device |
CN105701460A (en) * | 2016-01-07 | 2016-06-22 | 王跃明 | Video-based basketball goal detection method and device |
CN105975504A (en) * | 2016-04-28 | 2016-09-28 | 中国科学院计算技术研究所 | Recurrent neural network-based social network message burst detection method and system |
US10169656B2 (en) * | 2016-08-29 | 2019-01-01 | Nec Corporation | Video system using dual stage attention based recurrent neural network for future event prediction |
US20180203848A1 (en) * | 2017-01-17 | 2018-07-19 | Xerox Corporation | Author personality trait recognition from short texts with a deep compositional learning approach |
CN108108354A (en) * | 2017-06-18 | 2018-06-01 | 北京理工大学 | A kind of microblog users gender prediction's method based on deep learning |
CN107451118A (en) * | 2017-07-21 | 2017-12-08 | 西安电子科技大学 | Sentence-level sensibility classification method based on Weakly supervised deep learning |
CN109376784A (en) * | 2018-10-29 | 2019-02-22 | 四川大学 | A kind of personality prediction technique and personality prediction meanss |
CN109597891A (en) * | 2018-11-26 | 2019-04-09 | 重庆邮电大学 | Text emotion analysis method based on two-way length Memory Neural Networks in short-term |
CN109829154A (en) * | 2019-01-16 | 2019-05-31 | 中南民族大学 | Semantic-based personality prediction technique, user equipment, storage medium and device |
Non-Patent Citations (1)
Title |
---|
车丽萍: "《健康人格与自信》", 31 January 2012 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112487184A (en) * | 2020-11-26 | 2021-03-12 | 北京智源人工智能研究院 | User character judging method and device, memory and electronic equipment |
CN113268740A (en) * | 2021-05-27 | 2021-08-17 | 四川大学 | Input constraint completeness detection method of website system |
CN113221560A (en) * | 2021-05-31 | 2021-08-06 | 平安科技(深圳)有限公司 | Personality trait and emotion prediction method, personality trait and emotion prediction device, computer device, and medium |
Also Published As
Publication number | Publication date |
---|---|
CN110275953B (en) | 2021-11-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107291822B (en) | Problem classification model training method, classification method and device based on deep learning | |
Logeswaran et al. | Sentence ordering and coherence modeling using recurrent neural networks | |
CN106649603B (en) | Designated information pushing method based on emotion classification of webpage text data | |
CN111881262B (en) | Text emotion analysis method based on multi-channel neural network | |
CN106227721B (en) | Chinese Prosodic Hierarchy forecasting system | |
CN110287320A (en) | A kind of deep learning of combination attention mechanism is classified sentiment analysis model more | |
Lin et al. | A post-processing method for detecting unknown intent of dialogue system via pre-trained deep neural network classifier | |
CN110275953A (en) | Personality classification method and device | |
Anand Kumar et al. | A sequence labeling approach to morphological analyzer for tamil language | |
CN110377731A (en) | Complain text handling method, device, computer equipment and storage medium | |
CN110968725B (en) | Image content description information generation method, electronic device and storage medium | |
CN112632226B (en) | Semantic search method and device based on legal knowledge graph and electronic equipment | |
CN111985243A (en) | Emotion model training method, emotion analysis device and storage medium | |
Mattick et al. | SmartPatch: improving handwritten word imitation with patch discriminators | |
CN113704392A (en) | Method, device and equipment for extracting entity relationship in text and storage medium | |
Parvin et al. | Transformer-based local-global guidance for image captioning | |
CN111382243A (en) | Text category matching method, text category matching device and terminal | |
CN113723077A (en) | Sentence vector generation method and device based on bidirectional characterization model and computer equipment | |
CN115456176B (en) | Text matching method and system based on knowledge enhancement | |
Tashu et al. | Deep Learning Architecture for Automatic Essay Scoring | |
CN116644148A (en) | Keyword recognition method and device, electronic equipment and storage medium | |
US11880664B2 (en) | Identifying and transforming text difficult to understand by user | |
CN113792230B (en) | Service linking method, device, electronic equipment and storage medium | |
CN115495566A (en) | Dialog generation method and system for enhancing text features | |
CN113806475A (en) | Information reply method and device, electronic equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |