CN105206284B - Dredge the cyberchat method and system of adolescent psychology pressure - Google Patents
Dredge the cyberchat method and system of adolescent psychology pressure Download PDFInfo
- Publication number
- CN105206284B CN105206284B CN201510580511.XA CN201510580511A CN105206284B CN 105206284 B CN105206284 B CN 105206284B CN 201510580511 A CN201510580511 A CN 201510580511A CN 105206284 B CN105206284 B CN 105206284B
- Authority
- CN
- China
- Prior art keywords
- pressure
- word
- value
- sentence
- intermediate result
- 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.)
- Active
Links
Landscapes
- Machine Translation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention discloses a kind of cyberchat method and system for dredging adolescent psychology pressure, it is intended to by carrying out several times continuously dialogue response with teenager, some information actively and positively are provided, it guides them to think deeply with a kind of attitude of positive optimism, face the difficulty, achievees the purpose that alleviate adolescent psychology pressure.The method, comprising: S1, the first session sentence for obtaining user carry out Chinese word segmentation to the first session sentence, and establish the first dependency tree;S2, each node of the first dependency tree is labeled by the pressure dictionary pre-established, and the path length calculated between path length and pressure subtype node and mood word node between the pressure type node and mood word node after mark obtains intermediate result;S3, target intermediate result is selected from intermediate result, purpose pressure sensing results are extracted from target intermediate result;S4, target response sentence is obtained according to the purpose pressure sensing results, and feeds back to user.
Description
Technical field
The present invention relates to data analysis technique fields, and in particular to a kind of cyberchat side for dredging adolescent psychology pressure
Method and system.
Background technique
With the continuous quickening of social development, current people generally subject the psychological pressure from various aspects.Blueness is few
Year intelligence full maturity not yet, assurance and capacity of self-regulation to mood relative deficiency, excessive psychology compared with adult
Pressure often causes different degrees of injury to teen-age body and mind.Meanwhile the teenager in the phase of rebel is often unwilling
It is exchanged with other people, causes its pressure mood to be unable to get and timely dredge, cause more serious consequence.Correlative study table
Bright, the teenager of China 10%-30% has different degrees of psychological problems, and there are learning anxiety psychology by ninety percent or more student
The student of problem, nearly half nourishes the feared state of mind to examination.At the same time, with the rise of WEB2.0, network is more and more
Teenager provide exchange and express mood platform.Therefore, if a cyberchat tool can be designed and Implemented, by mutual
The dialogue that networking carries out intelligence with teenager is chatted, and is merged its content of microblog, is embodied in automatic identification teenager's session sentence
Pressure mood and source pressure provide the answer statement rich in hommization to dredge their psychological pressure, then for teenager
Healthy growth it is of great advantage, to entire society harmony health development cannot be neglected important function.
Since the 1990s, people begin to study and design automatically request-answering system.Automatically request-answering system carries out
One question answering process generally requires experience question text analysis, information retrieval, outcome evaluation and ranking, integrates the processes such as answer.
In terms of natural language question answering system, more famous system has Start, AskJeeves, AnswerBus etc..Start system
The problem of being a question answering system researched and developed by the Massachusetts Institute of Technology, history, geographical technology and culture etc. can be answered, still
It is also lacking in terms of the understanding to context.The major function of AskJeeves system is retrieval, passes through user's interaction and language
Justice analysis provides search result.AnswerBus is the multilingual automatically request-answering system of University of Michigan's research and development, it is in analysis user
After the natural language sentence of input, providing may be comprising 8 sentence and associated URL of answer.In addition, Microsoft
Encarta, the MULDER of University of Washington, LAM of National University of Singapore etc. are also the automatically request-answering systems of relative maturity.
In recent years, domestic many scientific research institutions and enterprise have also carried out phase to by the question answering system of dominant language of Chinese
Close research.2013, article " the A Chatterbot System for that Harbin Institute of Technology delivers on ACL
Answering Domain-specific Questions " describe the automatic question answering system around affective domain that they design
System.The problems in the problem of question answering system proposes user and database match, and filter out the problem of similarity maximum
And provide corresponding answer.
However intelligent Answer System has that three can not dredge applied to adolescent psychology pressure: 1) automatic question answering
System, which is often directed to, has the problem of objective or generally acknowledged answer, is unable to satisfy demand of the user when there is pressure mood;2) it " asks
Answer " mode lacks interaction: during pressure is dredged, the role of " quizmaster " and " answerer " should be in users and chat journey
It is suitably converted in sequence, to improve the interactivity for entirely dredging process;3) " question and answer " mode lacks the understanding to language environment:
The pouring out of user, chat process is a successional process, and each session sentence suffers from the context of context.If according to
The working principle of automatically request-answering system only merely replys each sentence respectively, then response effect is limited to depth.
Intelligent chat robots are a branches of automatically request-answering system, are generally basede on the side of pattern match to find one
Appropriate answer provides a relatively humanized answer to the natural language input of user, this answer may and not have
The characteristics of intellectual is more simulation person to person's chat conversations.The intelligent chat robots of comparative maturity have foreign countries at present
Eliza and Alice, domestic " small i ", " small figure " and " little Huang chicken " etc..Eliza was Joseph Weizenbaum in 1956
The chat robots of release are identified as first chat robots in the world.It has certain psychological knowledge, Neng Gouhe
User exchanges, but good continuity is short of between context.Alice is the chat robots based on experience,
By Richard doctor S.Wallace in nineteen ninety-five research and develop, its maximum feature is that artificial intelligence has been used to mark language
The template progress that AIML is sayed as Knowledge Description Language, for the natural sentence of user's input, after participle and in knowledge base
Match, obtains answer." small i " is a Chinese chat robots of Shanghai Incesoft Co., Ltd's independent research, is passed through
Development in 10 years, it has accumulated a large amount of relevant knowledge in multiple fields such as weather, map, translations, and supports speech recognition
Function is the Chinese chat robots of a comparative maturity." little Huang chicken " originates from Renren Network, hot rapidly later, is a
Pure chat robots.Its intelligence depends on huge backstage knowledge base, by instructional function, constantly learns user's answer
Keyword and answer and store.
Main problem of the intelligent chat robots in processing user's pressure mood is: 1) being full of negative feelings in face of user
The session sentence of thread, the response that a plurality of intelligence chat robots provide are not humanized enough;2) it is difficult to provide practical solution:
For example when user encounters school work pressure, chat tool should provide a user the guidance in some learning methods.And it is attempting
After the chat robots of several money prevalences, it is found that answer often contains when seeking advice from " What shall i do " similar problem to it
Its diction is pasted, practical opinion can not be provided.
In view of this, providing a kind of intelligent chat tool for the purpose of dredging adolescent psychology pressure for teenager becomes
Urgent problem to be solved.
Summary of the invention
The object of the present invention is to provide a kind of cyberchat method and systems for dredging adolescent psychology pressure, pass through
Continuously dialogue response is carried out several times with teenager, perceives the pressure mood and source pressure information of user, and according to user
Pressure mood and source pressure information, linked up, helped online in time, some information actively and positively are provided, him is guided
Thought deeply with a kind of attitude of positive optimism, faced the difficulty, achieve the purpose that alleviate adolescent psychology pressure.
For this purpose, on the one hand, the present invention proposes a kind of cyberchat method for dredging adolescent psychology pressure, comprising:
S1, obtain user the first session sentence, to the first session sentence carry out Chinese word segmentation, and establish first according to
Deposit relational tree, wherein the root node of first dependency tree is the centre word of the first session sentence, each child node
For the word that the progress Chinese word segmentation obtains, the dependence of each corresponding child node when indicating this;
S2, each node of first dependency tree is labeled by the pressure dictionary pre-established, and counted
The path length and pressure subtype node and mood word section between pressure type node and mood word node after calculating mark
Path length between point obtains intermediate result, wherein the pressure dictionary includes negative word and its interdependent positive mood
Pressure subtype word under word, pressure mood word, pressure type word, pressure type word and belong to corresponding to pressure subtype word
The descriptor of pressure subtype scope, the format of the intermediate result are (Stress, Category:lc, Subcategory:
ls), Stress has indicated whether pressure, and Category indicates pressure type word, lcIndicate pressure type word and mood word node it
Between path length, Subcategory indicate pressure subtype word, lsIt indicates between pressure subtype word and mood word node
Path length;
S3, target intermediate result is selected from the intermediate result, purpose pressure is extracted from the target intermediate result
Sensing results, wherein the format of the purpose pressure sensing results is (Stress, Category, Subcategory);
S4, target response sentence is obtained according to the purpose pressure sensing results, and feeds back to user.
On the other hand, the present invention proposes a kind of cyberchat system for dredging adolescent psychology pressure, comprising:
Unit is established, for obtaining the first session sentence of user, Chinese word segmentation is carried out to the first session sentence, and
Establish the first dependency tree, wherein the root node of first dependency tree is the centre word of the first session sentence,
Each child node is the word that the progress Chinese word segmentation obtains, the dependence of each corresponding child node when indicating this;
Computing unit marks each node of first dependency tree by the pressure dictionary pre-established
Note, and calculate the path length between the pressure type node and mood word node after mark and pressure subtype node and feelings
Path length between thread word node obtains intermediate result, wherein the pressure dictionary includes negative word and its is interdependent just
Pressure subtype word under face mood word, pressure mood word, pressure type word, pressure type word and belong to pressure subtype word institute
The descriptor of corresponding pressure subtype scope, the format of the intermediate result are (Stress, Category:lc,
Subcategory:ls), Stress has indicated whether pressure, and Category indicates pressure type word, lcIndicate pressure type word with
Path length between mood word node, Subcategory indicate pressure subtype word, lsIndicate pressure subtype word and mood
Path length between word node;
Extraction unit is mentioned from the target intermediate result for selecting target intermediate result from the intermediate result
Take purpose pressure to perceive as a result, wherein, the format of the purpose pressure sensing results be (Stress, Category,
Subcategory);
Acquiring unit for obtaining target response sentence according to the purpose pressure sensing results, and feeds back to user.
The cyberchat method and system that adolescent psychology pressure is dredged described in the embodiment of the present invention, by perceiving user
Mood word, pressure type word and the pressure subtype word and its dependence that session sentence is contained each time obtain purpose pressure
Sensing results, and corresponding target response sentence is obtained according to the purpose pressure sensing results, and feed back to user, Neng Goutong
It crosses and carries out several times continuously dialogue response with teenager, provide some information actively and positively for it, guide them with one kind
Actively optimistic attitude is thought deeply, is faced the difficulty, and achievees the purpose that alleviate adolescent psychology pressure.
Detailed description of the invention
Fig. 1 is the flow diagram for one embodiment of cyberchat method that the present invention dredges adolescent psychology pressure;
Fig. 2 is the pressure dictionary signal for another embodiment of cyberchat method that the present invention dredges adolescent psychology pressure
Figure;
Fig. 3 is that the present invention dredges the first dependency tree in the another embodiment of cyberchat method of adolescent psychology pressure
Schematic diagram;
Fig. 4 is that the present invention dredges normal direction overlapping between pedestrian in the another embodiment of cyberchat method of adolescent psychology pressure
And tangential displacement schematic diagram;
Fig. 5 is the frame structure schematic diagram for one embodiment of cyberchat system that the present invention dredges adolescent psychology pressure;
Fig. 6 is the architecture diagram for another embodiment of cyberchat system that the present invention dredges adolescent psychology pressure;
Fig. 7 is the log-in interface signal for the another embodiment of cyberchat system that the present invention dredges adolescent psychology pressure
Figure.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical solution in the embodiment of the present invention is explicitly described, it is clear that described embodiment is the present invention
A part of the embodiment, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not having
Every other embodiment obtained under the premise of creative work is made, shall fall within the protection scope of the present invention.
As shown in Figure 1, the present embodiment discloses a kind of cyberchat method for dredging adolescent psychology pressure, comprising:
S1, obtain user the first session sentence, to the first session sentence carry out Chinese word segmentation, and establish first according to
Deposit relational tree, wherein the root node of first dependency tree is the centre word of the first session sentence, each child node
For the word that the progress Chinese word segmentation obtains, the dependence of each corresponding child node when indicating this;
S2, each node of first dependency tree is labeled by the pressure dictionary pre-established, and counted
The path length and pressure subtype node and mood word section between pressure type node and mood word node after calculating mark
Path length between point obtains intermediate result, wherein the pressure dictionary includes negative word and its interdependent positive mood
Pressure subtype word under word, pressure mood word, pressure type word, pressure type word and belong to corresponding to pressure subtype word
The descriptor of pressure subtype scope, the format of the intermediate result are (Stress, Category:lc, Subcategory:
ls), Stress has indicated whether pressure, and Category indicates pressure type word, lcIndicate pressure type word and mood word node it
Between path length, Subcategory indicate pressure subtype word, lsIt indicates between pressure subtype word and mood word node
Path length;
S3, target intermediate result is selected from the intermediate result, purpose pressure is extracted from the target intermediate result
Sensing results, wherein the format of the purpose pressure sensing results is (Stress, Category, Subcategory);
S4, target response sentence is obtained according to the purpose pressure sensing results, and feeds back to user.
In the embodiment of the present invention, for the session sentence c of user's input, can define goal pressure sensing results is
(Stress, Category, Subcategory).In order to obtain the target of (Stress, Category, Subcategory) form
Pressure perception is as a result, can carry out Chinese word segmentation to sentence c first with language analysis tool LTP and be managed based on dependency grammar
By establishing a dependency tree.According to dependency grammar, sentence is an organized entirety, and the ingredient of composition is word, word with
Connection can be generated between word, connection constitutes the frame of sentence and expresses thoughts between these words.Therefore in the tree, each section
Point is the word carried out after Chinese word segmentation, and root node is the centre word in entire sentence, and each side indicates between two words
Dependence, the dependence between word and word have reacted the semantic relation between word and word.It is first based on this dependency tree
First according to the pressure dictionary having had built up (as shown in the table 1 of Fig. 2, including pressure mood dictionary and pressure type/subtype word
Library) each node is labeled.If there is the presence of pressure mood word or negative word and its interdependent front mood word
In the presence of then Stress value be 1, on the contrary it is then be 0.Next, analyzing the company between all pressure mood words and pressure type word
It connects, and obtains a series of shaped like (Stress, Category:lc, Subcategory:ls) intermediate result.Wherein Stress,
Whether Category, Subcategory are still respectively indicated pressure, pressure type and subtype, lcAnd lsPressure type node
With the path length (l of pressure subtype node and pressure mood word nodec/lsIt can indicate with value for 0 in pressure type section
Point/do not have path between pressure subtype node and mood word node).Finally, preferentially choosing lcIt is worth the smallest intermediate result to make
For target intermediate result, if lcBeing worth the smallest intermediate result has multiple, then chooses lsBe worth the smallest intermediate result as target among
As a result.But in some cases, lcAnd lsThe all the smallest intermediate result of value still has multiple, at this moment selection and historical pressures type
Identical and intermediate result with minimal path is as target intermediate result.And when historical information cannot help to select, then
One intermediate result with minimal path of random selection is as target intermediate result.To sum up, for the input meeting of user
Language sentence c will obtain a three-dimensional goal pressure sensing results (c.Stress, c.Category, c.Subcategory).
The cyberchat method that adolescent psychology pressure is dredged described in the embodiment of the present invention, each time by perception user
Mood word, pressure type word and the pressure subtype word and its dependence that session sentence is contained obtain purpose pressure perception knot
Fruit, and obtain corresponding target response sentence according to the purpose pressure sensing results, and feed back to user, can by with blueness
Teenager carries out several times continuously dialogue response, provides some information actively and positively for it, guides them with a kind of positive pleasure
The attitude of sight is thought deeply, is faced the difficulty, and achievees the purpose that alleviate adolescent psychology pressure.
Optionally, described logical in another embodiment for the cyberchat method that the present invention dredges adolescent psychology pressure
The pressure dictionary pre-established is crossed to be labeled each node of first dependency tree, comprising:
Judge in each node by searching for the pressure dictionary with the presence or absence of negative word and its interdependent front
Mood word or pressure mood word, and if it exists, then determine Stress value be have value corresponding to pressure, otherwise, it determines
The value of Stress is value corresponding to no pressure;
For each node, pressure type word belonging to the node is determined by searching for the pressure dictionary
Category and pressure subtype word Subcategory;
According to pressure type word Category belonging to the value of the Stress, each node and pressure subtype word
Subcategory obtains the value of (Stress, Category, Subcategory) format, and using obtain it is described (Stress,
Category, Subcategory) value of format is labeled each node.
Optionally, the present invention dredge adolescent psychology pressure cyberchat method another embodiment in, it is described from
Target intermediate result is selected in the intermediate result, comprising:
L is selected from the intermediate resultcThe smallest intermediate result;
Judge the lcWhether the quantity of the smallest intermediate result is 1, if 1, then by the lcThe smallest intermediate result is made
For the purpose intermediate result, or
If not 1, then from the lcL is selected in the smallest intermediate resultcAnd lsThe smallest intermediate result judges the lc
And lsWhether the quantity of the smallest intermediate result is 1, if 1, then by the lcAnd lsThe smallest intermediate result is as the purpose
Intermediate result, or
If the lcAnd lsThe quantity of the smallest intermediate result is not 1, then judges the lcAnd lsIn the smallest intermediate result
With the presence or absence of intermediate result identical with the pressure type word of historical purpose pressure sensing results, and if it exists, then from the lcAnd ls
It is chosen in one in intermediate result identical with the pressure type word of historical purpose pressure sensing results in the smallest intermediate result
Between result as the purpose intermediate result, otherwise, if it does not exist, then from the lcAnd lsOne is chosen in the smallest intermediate result
A intermediate result is as the purpose intermediate result.
In the embodiment of the present invention, with the session sentence-of user's input, " summation friend quarrels after the new term begins, good agitation!" be
Example is illustrated the acquisition process of purpose pressure sensing results of the present invention.It is " tired as shown in the dependency tree of Fig. 3
It is hot-tempered " it is labeled as pressure mood word, " starting to school " is labeled as school work pressure type word, and " friend " is labeled as the pressure of the friend in interpersonal relationships
Power subtype word, " quarrel " are labeled as interpersonal relationships pressure type word.Due to there is " agitation " this pressure mood word,
Stress value is 1.And the relationship between pressure type and subtype word and pressure mood word is analyzed, available two centres
As a result: (1, school work: 3, general general:0), (1, interpersonal relationships: 1, friend: 3).Based on shortest path principle, can choose
(1, interpersonal relationships, friend) goal pressure sensing results are used as.
However, the chat sentence of user is all very brief in actual conditions, the three-dimensional pressure knot perceived out by above method
The case where usually will appear c.Stress=0 or c.Category=null or c.Subcategory=null in fruit, i.e. user
Read statement there is no the specific mood words or pressure type word for describing pressure condition.Based on pressure mood in chat process
Duration, constantly record and update user historical pressures state (h.Stress, h.Category, h.Subcategory)
(by perceiving user's pressure state result in pervious read statement), and whereby to the pressure state currently perceived
Missing values in (c.Stress, c.Category, c.Subcategory) are supplemented, and following situation is specifically divided into:
(1) as c.Stress=0 and when h.Stress=0, if c.Category=h.Category,
C.Subcategory=h.Subcategory then enables c.Stress=1;
(2) as c.Stress=1 while h.Stress=1, if c.Category=null, c.Subcategory=
Null or c.Category=h.Category, c.Subcategory=null, enable c.Category=h.Category,
C.Subcategory=h.Subcategory.
Final (c.Stress, c.Category, c.Subcategory) is recorded as to new historical pressures state
(h.Stress, h.Category, h.Subcategory).
Optionally, in another embodiment for the cyberchat method that the present invention dredges adolescent psychology pressure, described
After selecting target intermediate result in the intermediate result, purpose pressure sensing results are extracted from the target intermediate result
Before, further includes:
The value of the value and the Stress of historical purpose pressure sensing results that judge the Stress of the target intermediate result is
No is to have value corresponding to pressure, if the value of the Stress of the historical purpose pressure sensing results is to have corresponding to pressure
Value, the value of the Stress of the target intermediate result are value corresponding to no pressure, then judge the historical purpose pressure perception
As a result the value of the Category of the value of Category and the target intermediate result and historical purpose pressure perception are tied
Whether the value of the Subcategory of fruit and the value of the Subcategory of the target intermediate result are equal, if equal, by institute
State the Stress of target intermediate result value be updated to the historical purpose pressure sensing results Stress value, or
If the value of the Stress of the value of the Stress of the historical purpose pressure sensing results and the target intermediate result
To there is value corresponding to pressure, then judge the Category of the target intermediate result value and Subcategory value whether
The value of the Category of the target intermediate result is then updated to the historical purpose pressure and is perceived if null value for null value
The value of the Subcategory of the target intermediate result is updated to the historical purpose pressure by the value of Category as a result
The value of the Subcategory of sensing results, otherwise, if the value of the Category of the target intermediate result is not null value,
The value of Subcategory is null value, then judges the value and the target of the Category of the historical purpose pressure sensing results
Whether the value of the Category of intermediate result is equal, if equal, more by the value of the Subcategory of the target intermediate result
It is newly the value of the Subcategory of the historical purpose pressure sensing results.
Optionally, in another embodiment for the cyberchat method that the present invention dredges adolescent psychology pressure, described
Target response sentence is obtained according to the purpose pressure sensing results, comprising:
Determine the type of the first session sentence, wherein the type includes interrogative sentence and non-interrogative sentence;
Judge the first session sentence type whether be non-interrogative sentence and purpose pressure sensing results Stress
Value whether be have value corresponding to pressure, if the type of the first session sentence be non-interrogative sentence, purpose pressure perception knot
The value of the Stress of fruit be have value corresponding to pressure, if the first session sentence type be interrogative sentence, purpose pressure
The value of the Stress of power sensing results is value corresponding to no pressure, then searches the first knowledge base pre-established, is obtained described
Answer statement corresponding to the value of Category and the value of Subcategory of purpose pressure sensing results, and lookup is obtained
Answer statement as the target response sentence, wherein preserved in first knowledge base a plurality of value by Category,
The record of value and the answer statement composition of Subcategory, or
If the type of the first session sentence is non-interrogative sentence, the value of the Stress of purpose pressure sensing results is no pressure
Value corresponding to power then chooses the target response sentence from the second knowledge base, wherein preserves in second knowledge base
A plurality of answer statement, or
If the type of the first session sentence is interrogative sentence, the value of the Stress of purpose pressure sensing results is to have pressure
Corresponding value then obtains the target response sentence from third knowledge base, wherein includes multiple in the third knowledge base
Problem sentence and corresponding answer.
In the embodiment of the present invention, type and goal pressure sensing results based on user's read statement (c.Stress,
C.Category, c.Subcategory), the acknowledgment strategy of this chat tool is as shown in Figure 4.
Situation one: pressure perceives (c.Stress=1)
(1) for declarative sentence and confirmative question, the present invention can the selection from the first knowledge base (can be local knowledge base)
One suitable response returns to user.Store a large amount of energy for being directed to different pressures type and subtype in local knowledge base
Comfort is enough played, is encouraged, or the answer statement of guidance user.
(2) for interrogative sentence, this chat tool can search one from third knowledge base (Chinese maximum Ask-Answer Community)
Most suitable answer statement returns to user, and specific working principle is as follows:
Enable QuAnd QbThe problem of respectively user inputs sentence and the maximally related problem language searched from third knowledge base
Sentence.As MatchDegree (Qu,Qb) it is greater than a threshold value (i.e. problem sentence QuWith maximally related problem sentence QbMatching degree),
And the answer of the problem by user adopt or at least one people approve of when, present invention meeting according to whether the person of being asked adopt with
And it approves of number to choose best answer and returns to user.Otherwise, the present invention is general back to selection one in local knowledge base
Encouragement language or a joke return to user.
Wherein MatchDegree (Qu,Qb) it is based on QuAnd QbDependency tree in effective node to calculating.It enables
Effective node is to for n (n1,n2), n1For central node, that is, root node in dependency tree, n2For with n1Node have directly according to
The node of relationship is deposited, it can be a verb, noun either adjective.Assuming that two effective nodes are to for u=(u1,u2)
With b=(b1,b2), then it is as follows can to define similarity sim (u, b):
Enable E (Qu) and E (Qb) it is respectively QuAnd QbDependence tree species effective node to collection.In view of the language of node
Adopted similitude can replace original with the justice former (can obtain from Hownet Chinese vocabulary bank Hownet) of word in each node simultaneously
The word of beginning is to guarantee to obtain more accurate similarity.Thus to obtain effective node to integrating as E*(Qu) and E*(Qb).It is possible thereby to
Define MatchDegree (Qu,Qb) are as follows:
MatchDegree(Qu,Qb)=α SIM (E (Qu),E(Qb))+(1-α)SIM(E*(Qu),E*(Qb)),
Situation two: no pressure perceives (c.Stress=0)
(1) for non-interrogative sentence, the present invention will call the API of the chat robots simsimi of open source, the API to receive use
One chat input by sentence at family returns to answer statement;
(2) identical as situation one (2) for interrogative sentence.
Optionally, the present invention dredge adolescent psychology pressure cyberchat method another embodiment in, it is described from
The target response sentence is obtained in third knowledge base, comprising:
It is obtained and the maximally related problem sentence of the first session sentence from third knowledge base, wherein the third is known
Know in library further includes the information whether multiple corresponding answers of problem sentence are adopted and the quantity that corresponding answer is approved of letter
Breath;
The matching degree for calculating the first session sentence Yu maximally related problem sentence, judges whether the matching degree is greater than
Preset numerical value then judges whether the answer of the maximally related problem sentence is adopted if more than the numerical value, if by adopting
It receives, then chooses the answer being adopted for approving of quantity most from the answer being adopted as the target response sentence, alternatively,
If no more than the numerical value or not being adopted, a drum is randomly selected from the encouragement language that first knowledge base saves
Language is encouraged as the target response sentence.
Optionally, in another embodiment for the cyberchat method that the present invention dredges adolescent psychology pressure, the meter
Calculate the matching degree of the first session sentence Yu maximally related problem sentence, comprising:
To the maximally related problem sentence QbChinese word segmentation is carried out, and establishes the second dependency tree;
Determine effective node of first dependency tree to u=(u1,u2) and the second dependency tree effective section
Point is to b=(b1,b2), wherein u=(u1,u2), u1For the root node of first dependency tree, u2For with u1With direct
The node of dependence, b1For the root node of second dependency tree, b2For with b1Node with direct dependence;
Calculate the first session sentence QuWith maximally related problem sentence QbMatching degree,
Calculation formula is MatchDegree (Qu,Qb)=α SIM (E (Qu),E(Qb))+(1-α)SIM(E*(Qu),E*(Qb)),
Wherein, MatchDegree (Qu,Qb) it is the first session sentence QuWith maximally related problem sentence QbMatching degree, α (α ∈
(0,1)) it is constant, E (Qu) be first dependency tree effective node pair set, E (Qb) it is the described second interdependent pass
It is the set of effective node pair of tree, E*(Qu) it is that will be obtained after each node of first dependency tree former substitution of its justice
The set of effective node pair of the dependency tree arrived, E*(Qb) it is by its justice of each node of second dependency tree
The set of effective node pair of the dependency tree obtained after original substitution, u*For E*(Qu) element, b*For E*
(Qb) element,
Optionally, in another embodiment for the cyberchat method that the present invention dredges adolescent psychology pressure, described
Before the first session sentence for obtaining user, further includes:
The direct log-in instruction of user is received, and the first answer statement is fed back into user, or
The log-in instruction of user is received, and when user authorizes connection microblogging, the original microblogging of user is obtained, according to described
Original microblogging judges whether user has pressure mood, if so, the second answer statement is then fed back to user, otherwise, then by third
Answer statement feeds back to user.
Optionally, in another embodiment for the cyberchat method that the present invention dredges adolescent psychology pressure, described
Before the first session sentence for obtaining user, further includes:
After first answer statement, the second answer statement or third answer statement are fed back to user, obtain
The second session sentence of user carries out pressure mood to the second session sentence and pressure type detects;
If detecting pressure mood in the second session sentence, pressure type is not detected, then to the original microblogging
Pressure detecting is carried out, and the 4th answer statement is fed back to by user based on the pressure sensing result, or
If detecting pressure mood and pressure type in the second session sentence or not detecting pressure mood, hold
Row step S1.
In addition, text emotion analysis is to be analyzed subjective texts, concluded, the process of its emotional characteristics is obtained.It is logical
The emotion information extracted in text, the description object classified to emotion information, analyze emotion vocabulary are crossed to judge phrase, sentence
Sentiment orientation in son even entire article.Dependency grammar is the " structure write by French linguistics L.Tesniere at it
Syntax basis " in propose.Sentiment analysis work more mature, existing literature (Qi Li, Yuanyuan based on microblog
Xue,Jia Jia,Ling Feng:Helping Teenagers Relieve Psychological Pressures:A
Micro-blog Based System.[C]EDBT 2014:660-663;W.Che,Z.Li,Y.Li,Y.Guo,B.Qin,and
T.Liu,“Multilingual dependency-based syntactic and semantic parsing,”in
Proc.of CoNLL,2009,pp.49–54.;[15]W.Che,Z.Li,and T.Liu,"LTP:a chinese language
Technology platform, " in Proc.of Coling, 2010, pp.13-16.) it is directed to the microblogging that teenager is issued,
Semantic analysis has been carried out to typical psychological pressure therein, based on the method for microblogging feature extraction adolescent psychology Pressure Analysis,
And more deep investigation has been carried out to adolescent psychology problem by PSS-14 psychological research questionnaire, it proposes based on microblogging
The psychological pressure of "@" mode dredges tool.The present invention carries out pressure detecting to the original microblogging, and is based on the pressure detecting
As a result the embodiment that the 4th answer statement feeds back to user can specifically be used into the above-mentioned prior art.
As shown in figure 5, the present embodiment discloses a kind of cyberchat system for dredging adolescent psychology pressure, comprising:
Unit 1 is established, for obtaining the first session sentence of user, Chinese word segmentation is carried out to the first session sentence,
And establish the first dependency tree, wherein the root node of first dependency tree is the center of the first session sentence
Word, each child node are the word that the progress Chinese word segmentation obtains, the interdependent pass of each corresponding child node when indicating this
System;
Computing unit 2 marks each node of first dependency tree by the pressure dictionary pre-established
Note, and calculate the path length between the pressure type node and mood word node after mark and pressure subtype node and feelings
Path length between thread word node obtains intermediate result, wherein the pressure dictionary includes negative word and its is interdependent just
Pressure subtype word under face mood word, pressure mood word, pressure type word, pressure type word and belong to pressure subtype word institute
The descriptor of corresponding pressure subtype scope, the format of the intermediate result are (Stress, Category:lc,
Subcategory:ls), Stress has indicated whether pressure, and Category indicates pressure type word, lcIndicate pressure type word with
Path length between mood word node, Subcategory indicate pressure subtype word, lsIndicate pressure subtype word and mood
Path length between word node;
Extraction unit 3 is mentioned from the target intermediate result for selecting target intermediate result from the intermediate result
Take purpose pressure to perceive as a result, wherein, the format of the purpose pressure sensing results be (Stress, Category,
Subcategory);
Acquiring unit 4 for obtaining target response sentence according to the purpose pressure sensing results, and feeds back to user.
The cyberchat system that adolescent psychology pressure is dredged described in the embodiment of the present invention, each time by perception user
Mood word, pressure type word and the pressure subtype word and its dependence that session sentence is contained obtain purpose pressure perception knot
Fruit, and obtain corresponding target response sentence according to the purpose pressure sensing results, and feed back to user, can by with blueness
Teenager carries out several times continuously dialogue response, provides some information actively and positively for it, guides them with a kind of positive pleasure
The attitude of sight is thought deeply, is faced the difficulty, and achievees the purpose that alleviate adolescent psychology pressure.
The present invention constructs the solution of a cyberchat for detecting and alleviating adolescent psychology pressure, such as schemes
Shown in 6.This system specifically includes following three big modules: chat manager, pressure perceptron and response generator.
For chat manager, when user's login system, system will be prompted to user and connect microblogging (login interface such as Fig. 7 institute
Show).If user's with no authorized microblogging links, it is directly entered chat interface and starts conventional chat program;When user's selection is awarded
After the connection of power system microblogging, webpage jumps to chat page.Meanwhile system will access microblogging relevant interface, obtain user publication
Microblogging.After obtaining the original microblogging of user, by analyzing whether the original microblogging that user delivers recently has pressure mood,
Adjustment is sent to the opening remarks sentence of user.If system unauthorized connects user's microblogging, or does not examine from user's microblogging
Pressure mood is measured, system will be with the greeting user of friend's formula: " hello!" and start conventional chat program., whereas if inspection
Pressure mood is measured, system will greet user: " hello!Recently how is everything going? " or " all are good and sound? ".What if user replied
Chat sentence detects pressure mood but no pressure type, and then the pressure sensing result based on microblogging tastes user to system
The inquiry of examination property advances to conventional chat program to guide user more specifically to describe his pressure condition later;If with
The chat sentence that family is replied detects pressure mood and pressure type simultaneously, or does not detect pressure mood, then system
Then the read statement current to user of chat program routinely carries out response.
Pressure perceptron can specifically be set in order to which the mood and desired answer of user's expression is more completely understood
Meter sentence type identification module and content of the sentence analysis module are analyzed come the chat sentence to user.Wherein sentence type is known
The chat sentence of user can be divided into interrogative sentence, confirmative question and declarative sentence by other module.For example, how " me could be improved
Mathematics achievement? " or " I should be what if? " etc. being interrogative sentence, which, which may expect to obtain, some is objectively built
View or pep talk language.And for another example " they should not be understood me? " confirmative question then to express user's one kind stronger
Mood, he may is that expectation obtains the answer pacified.On this basis, content of the sentence analysis module further analyzes user's input
The content of sentence is thus to perceive the specific pressure condition of user (type and subtype including pressure).In view of user chats
The continuity of process, the module record the pressure state of user's history simultaneously, and obtain more accurately by historical information
Current pressure sensing results.
For response generator, the statement type result generated based on pressure perceptron analysis and pressure state are as a result, answer
Selecting module is answered by one conjunction of selection from the library chat robots simsimi of local knowledge base, third knowledge base and open source
Suitable answer returns to user.The module allows user to express the pain of heart or shifts the note of emotion by the interaction with user
Meaning power is to help user to dredge pressure.Response management module is mainly responsible for creation and management acknowledgment strategy and Indigenous knowledge
Library.The module has collected from Internet resources abundant largely there is pressure to dredge the answer statement of function and be stored in local
Knowledge base, to provide response resource abundant for response selecting module.
For the present invention, following user's experiment has been carried out:
Student of 10 ages between 15 years old to 22 years old has been invited to participate in from middle school and university one month by a definite date
User experiment in.Select this 10 people be because teen-age main body be all student and also participate in experiment 10 people also just meeting with
By more or less pressure from various aspects.Experiment at the beginning, informs that experiment this chat tool of participant can help
They dredge psychological pressure, they can chat with our chat tool at any time.In test process, participant is tested
Do following three things: (1) every filling in a pressure consciousness scale (CPSS-14) week, (pressure consciousness scale is the heart
It is widely used in measuring the other scale of people's pressure stage in Neo-Confucianism) record their pressure changings.(2) every to complete primary meeting
After words process, effect is dredged for pressure and gives a scoring.(3) judge each chat sentence pressure perception the result is that
It is no correct.
For each session, each read statement of user and its corresponding pressure sensing results are returned into user,
Judge that pressure perceives the result is that no correct by user.The accuracy of pressure perception algorithm is evaluated using accuracy rate and recall rate.
78.34% and 76.12% are reached for the precision ratio and recall ratio of the perception of pressure existence.Perception for pressure type,
The recall rate and other kinds of accuracy rate of emotion and human communication type are lower, this is because many emotions and human communication
The pressure of type is perceived as other pressure types.In addition, for school work pressure most generally existing in teenager, this chat work
Tool has preferable perception accuracy, and precision ratio and recall ratio have respectively reached 86.2% and 75%.
For the validity that the chat tool pressure is dredged, user after session each time will for pressure dredge effect to
Giving a score, (1 point: mood is worse, and 2 points: mood does not change, and 3 points: mood is felt a little to improve, and 4 points: mood is felt
Turn, 5 points: mood has biggish improvement).There is the score of 60% or more session to be above 2 points, i.e., 60% or more session is all
User mood can be allowed to improve.This chat tool dredges the pressure of learning type that effect is preferable, this is because the pressure of learning type
Power sensing results are more accurate.
Although the embodiments of the invention are described in conjunction with the attached drawings, but those skilled in the art can not depart from this hair
Various modifications and variations are made in the case where bright spirit and scope, such modifications and variations are each fallen within by appended claims
Within limited range.
Claims (10)
1. a kind of cyberchat method for dredging adolescent psychology pressure characterized by comprising
S1, the first session sentence for obtaining user carry out Chinese word segmentation to the first session sentence, and establish the first interdependent pass
System tree, wherein the root node of first dependency tree is the centre word of the first session sentence, and each child node is institute
It states and carries out the word that Chinese word segmentation obtains, the dependence of each corresponding child node when indicating this;
S2, each node of first dependency tree is labeled by the pressure dictionary pre-established, and calculates mark
Path length and pressure subtype node between pressure type node after note and mood word node and mood word node it
Between path length obtain intermediate result, wherein the pressure dictionary includes negative word and its interdependent front mood word, pressure
Pressure subtype word under power mood word, pressure type word, pressure type word and belong to pressure corresponding to pressure subtype word
The descriptor of subtype scope, the format of the intermediate result are (Stress, Category:lc, Subcategory:ls),
Stress has indicated whether pressure, and Category indicates pressure type word, lcIt indicates between pressure type word and mood word node
Path length, Subcategory indicate pressure subtype word, lsIndicate the road between pressure subtype word and mood word node
Electrical path length;
S3, target intermediate result is selected from the intermediate result, the perception of purpose pressure is extracted from the target intermediate result
As a result, wherein the format of the purpose pressure sensing results is (Stress, Category, Subcategory);
S4, target response sentence is obtained according to the purpose pressure sensing results, and feeds back to user.
2. the cyberchat method according to claim 1 for dredging adolescent psychology pressure, which is characterized in that described to pass through
The pressure dictionary pre-established is labeled each node of first dependency tree, comprising:
Judge in each node by searching for the pressure dictionary with the presence or absence of negative word and its interdependent positive mood
Word or pressure mood word, and if it exists, the value for then determining Stress is to have value corresponding to pressure, otherwise, it determines Stress
Value is value corresponding to no pressure;
For each node, by searching for the pressure dictionary determine pressure type word Category belonging to the node and
Pressure subtype word Subcategory;
According to pressure type word Category belonging to the value of the Stress, each node and pressure subtype word
Subcategory obtains the value of (Stress, Category, Subcategory) format, and using obtain it is described (Stress,
Category, Subcategory) value of format is labeled each node.
3. the cyberchat method according to claim 1 for dredging adolescent psychology pressure, which is characterized in that described from institute
It states and selects target intermediate result in intermediate result, comprising:
L is selected from the intermediate resultcThe smallest intermediate result;
Judge the lcWhether the quantity of the smallest intermediate result is 1, if 1, then by the lcThe smallest intermediate result is as institute
Purpose intermediate result is stated, or
If not 1, then from the lcL is selected in the smallest intermediate resultcAnd lsThe smallest intermediate result judges the lcAnd lsMost
Whether the quantity of small intermediate result is 1, if 1, then by the lcAnd lsThe smallest intermediate result is as knot among the purpose
Fruit, or
If the lcAnd lsThe quantity of the smallest intermediate result is not 1, then judges the lcAnd lsIn the smallest intermediate result whether
In the presence of intermediate result identical with the pressure type word of historical purpose pressure sensing results, and if it exists, then from the lcAnd lsIt is minimum
Intermediate result in an intermediate knot is chosen in intermediate result identical with the pressure type word of historical purpose pressure sensing results
Fruit is as the purpose intermediate result, otherwise, if it does not exist, then from the lcAnd lsIt is chosen in one in the smallest intermediate result
Between result as the purpose intermediate result.
4. the cyberchat method according to claim 1 for dredging adolescent psychology pressure, which is characterized in that it is described from
After selecting target intermediate result in the intermediate result, from the target intermediate result extract purpose pressure sensing results it
Before, further includes:
Judge the value of the Stress of the target intermediate result and the Stress of historical purpose pressure sensing results value whether be
There is value corresponding to pressure, if the value of the Stress of the historical purpose pressure sensing results is to have value corresponding to pressure, institute
The value for stating the Stress of target intermediate result is value corresponding to no pressure, then judges the historical purpose pressure sensing results
The value of the Category of the value of Category and the target intermediate result and the historical purpose pressure sensing results
Whether the value of Subcategory and the value of the Subcategory of the target intermediate result are equal, if equal, by the mesh
The value of the Stress of mark intermediate result is updated to the value of the Stress of the historical purpose pressure sensing results, or
If the value of the Stress of the historical purpose pressure sensing results and the value of the Stress of the target intermediate result are to have
Whether the value of value corresponding to pressure, the then value and Subcategory that judge the Category of the target intermediate result is empty
Value, if null value, is then updated to the historical purpose pressure sensing results for the value of the Category of the target intermediate result
Category value, the value of the Subcategory of the target intermediate result is updated to the historical purpose pressure and is perceived
As a result the value of Subcategory, otherwise, if the value of the Category of the target intermediate result is not null value,
The value of Subcategory is null value, then judges the value and the target of the Category of the historical purpose pressure sensing results
Whether the value of the Category of intermediate result is equal, if equal, more by the value of the Subcategory of the target intermediate result
It is newly the value of the Subcategory of the historical purpose pressure sensing results.
5. the cyberchat method according to claim 1 for dredging adolescent psychology pressure, which is characterized in that the basis
The purpose pressure sensing results obtain target response sentence, comprising:
Determine the type of the first session sentence, wherein the type includes interrogative sentence and non-interrogative sentence;
Judge the first session sentence type whether be non-interrogative sentence and purpose pressure sensing results Stress value
It whether is to have value corresponding to pressure, if the type of the first session sentence is non-interrogative sentence, purpose pressure sensing results
The value of Stress be have value corresponding to pressure, if the first session sentence type be interrogative sentence, purpose feeling of stress
The value for knowing the Stress of result is value corresponding to no pressure, then searches the first knowledge base pre-established, obtain the purpose
Answer statement corresponding to the value of the Category of pressure sensing results and the value of Subcategory, and answered what lookup obtained
Sentence is answered as the target response sentence, wherein preserved in first knowledge base a plurality of value by Category,
The record of value and the answer statement composition of Subcategory, or
If the type of the first session sentence is non-interrogative sentence, the value of the Stress of purpose pressure sensing results is no pressure institute
Corresponding value then chooses the target response sentence from the second knowledge base, wherein preserves in second knowledge base a plurality of
Answer statement, or
If the type of the first session sentence is interrogative sentence, the value of the Stress of purpose pressure sensing results is right to there is pressure
The value answered then obtains the target response sentence from third knowledge base, wherein includes multiple problems in the third knowledge base
Sentence and corresponding answer.
6. the cyberchat method according to claim 5 for dredging adolescent psychology pressure, which is characterized in that described from
The target response sentence is obtained in three knowledge bases, comprising:
It is obtained and the maximally related problem sentence of the first session sentence from third knowledge base, wherein the third knowledge base
In further include the information whether multiple corresponding answers of problem sentence are adopted and the quantity information that corresponding answer is approved of;
It is default to judge whether the matching degree is greater than for the matching degree for calculating the first session sentence Yu maximally related problem sentence
Numerical value then judge whether the answer of the maximally related problem sentence is adopted if more than the numerical value, if being adopted,
The answer being adopted for approving of quantity most is chosen from the answer being adopted as the target response sentence, alternatively, if not
It greater than the numerical value or is not adopted, then randomly selects an encouragement words from the encouragement language that first knowledge base saves
Language is as the target response sentence.
7. the cyberchat method according to claim 6 for dredging adolescent psychology pressure, which is characterized in that the calculating
The matching degree of the first session sentence and maximally related problem sentence, comprising:
To the maximally related problem sentence QbChinese word segmentation is carried out, and establishes the second dependency tree;
Determine effective node of first dependency tree to u=(u1,u2) and the second dependency tree effective node pair
B=(b1,b2), wherein u=(u1,u2), u1For the root node of first dependency tree, u2For with u1With directly interdependent
The node of relationship, b1For the root node of second dependency tree, b2For with b1Node with direct dependence;
Calculate the first session sentence QuWith maximally related problem sentence QbMatching degree,
Calculation formula is MatchDegree (Qu,Qb)=α SIM (E (Qu),E(Qb))+(1-α)SIM(E*(Qu),E*(Qb)),
In, MatchDegree (Qu,Qb) it is the first session sentence QuWith maximally related problem sentence QbMatching degree, α is constant,
α ∈ (0,1), E (Qu) be first dependency tree effective node pair set, E (Qb) it is second dependency tree
Effective node pair set, E*(Qu) it is that will be obtained after each node of first dependency tree former substitution of its justice
The set of effective node pair of dependency tree, E*(Qb) it is to replace each node of second dependency tree with its adopted original
The set of effective node pair of the dependency tree obtained after generation, u*For E*(Qu) element, b*For E*
(Qb) element,
8. the cyberchat method according to claim 1 for dredging adolescent psychology pressure, which is characterized in that obtained described
Before taking the first session sentence at family, further includes:
The direct log-in instruction of user is received, and the first answer statement is fed back into user, or
The log-in instruction of user is received, and when user authorizes connection microblogging, the original microblogging of user is obtained, according to described original
Microblogging judges whether user has pressure mood, if so, the second answer statement is then fed back to user, otherwise, then by third response
Sentence feeds back to user.
9. the cyberchat method according to claim 8 for dredging adolescent psychology pressure, which is characterized in that obtained described
Before taking the first session sentence at family, further includes:
After first answer statement, the second answer statement or third answer statement are fed back to user, user is obtained
The second session sentence, pressure mood and pressure type are carried out to the second session sentence and detected;
If detecting pressure mood in the second session sentence, pressure type is not detected, then the original microblogging is carried out
Pressure detecting, and the 4th answer statement is fed back to by user based on the pressure sensing result, or
If detecting pressure mood and pressure type in the second session sentence or not detecting pressure mood, step is executed
Rapid S1.
10. a kind of cyberchat device for dredging adolescent psychology pressure characterized by comprising
Unit is established, for obtaining the first session sentence of user, Chinese word segmentation is carried out to the first session sentence, and establish
First dependency tree, wherein the root node of first dependency tree is the centre word of the first session sentence, each
Child node is the word that the progress Chinese word segmentation obtains, the dependence of each corresponding child node when indicating this;
Computing unit is labeled each node of first dependency tree by the pressure dictionary pre-established, and
The path length and pressure subtype node and mood word between pressure type node and mood word node after calculating mark
Path length between node obtains intermediate result, wherein the pressure dictionary includes negative word and its interdependent positive feelings
Pressure subtype word under thread word, pressure mood word, pressure type word, pressure type word and belong to corresponding to pressure subtype word
Pressure subtype scope descriptor, the format of the intermediate result is (Stress, Category:lc, Subcategory:
ls), Stress has indicated whether pressure, and Category indicates pressure type word, lcIndicate pressure type word and mood word node it
Between path length, Subcategory indicate pressure subtype word, lsIt indicates between pressure subtype word and mood word node
Path length;
Extraction unit extracts mesh from the target intermediate result for selecting target intermediate result from the intermediate result
Pressure perception as a result, wherein, the format of the purpose pressure sensing results be (Stress, Category,
Subcategory);
Acquiring unit for obtaining target response sentence according to the purpose pressure sensing results, and feeds back to user.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510580511.XA CN105206284B (en) | 2015-09-11 | 2015-09-11 | Dredge the cyberchat method and system of adolescent psychology pressure |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510580511.XA CN105206284B (en) | 2015-09-11 | 2015-09-11 | Dredge the cyberchat method and system of adolescent psychology pressure |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105206284A CN105206284A (en) | 2015-12-30 |
CN105206284B true CN105206284B (en) | 2019-06-18 |
Family
ID=54953913
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510580511.XA Active CN105206284B (en) | 2015-09-11 | 2015-09-11 | Dredge the cyberchat method and system of adolescent psychology pressure |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105206284B (en) |
Families Citing this family (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106202052A (en) * | 2016-07-20 | 2016-12-07 | 清华大学 | Adolescent psychology pressure range and pressure source event perception method and device thereof |
US9947319B1 (en) * | 2016-09-27 | 2018-04-17 | Google Llc | Forming chatbot output based on user state |
CN106503786B (en) * | 2016-10-11 | 2020-06-26 | 北京光年无限科技有限公司 | Multi-modal interaction method and device for intelligent robot |
WO2018205224A1 (en) * | 2017-05-11 | 2018-11-15 | Microsoft Technology Licensing, Llc | Assisting psychological cure in automated chatting |
CN108255804A (en) * | 2017-09-25 | 2018-07-06 | 上海四宸软件技术有限公司 | A kind of communication artificial intelligence system and its language processing method |
CN107943974A (en) * | 2017-11-28 | 2018-04-20 | 合肥工业大学 | Consider the automatic session method and system of emotion |
CN108052605A (en) * | 2017-12-13 | 2018-05-18 | 广州佰聆数据股份有限公司 | A kind of intelligent Answer System based on client feature library |
KR102047010B1 (en) * | 2017-12-21 | 2019-11-20 | 주식회사 카카오 | Server, device and method for providing instant messeging service by using relay chatbot |
CN108511071A (en) * | 2018-04-18 | 2018-09-07 | 广东惠禾科技发展有限公司 | Mental health evaluation method and device |
CN108763355B (en) * | 2018-05-16 | 2021-12-17 | 深圳市三宝创新智能有限公司 | User-based intelligent robot interactive data processing system and method |
CN108877786A (en) * | 2018-06-29 | 2018-11-23 | 恒信东方文化股份有限公司 | Instruction identification method |
CN110851560B (en) * | 2018-07-27 | 2023-03-10 | 杭州海康威视数字技术股份有限公司 | Information retrieval method, device and equipment |
CN109243582A (en) * | 2018-09-19 | 2019-01-18 | 江苏金惠甫山软件科技有限公司 | The human-computer interaction motion management method and system of knowledge based graphical spectrum technology |
CN111261173A (en) * | 2020-01-10 | 2020-06-09 | 珠海格力电器股份有限公司 | Electric appliance control method and device, storage medium and electric appliance |
CN112989037B (en) * | 2021-02-05 | 2023-05-23 | 浙江连信科技有限公司 | Information processing method and device for identifying occupational pressure sources |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101075435A (en) * | 2007-04-19 | 2007-11-21 | 深圳先进技术研究院 | Intelligent chatting system and its realizing method |
CN104050256A (en) * | 2014-06-13 | 2014-09-17 | 西安蒜泥电子科技有限责任公司 | Initiative study-based questioning and answering method and questioning and answering system adopting initiative study-based questioning and answering method |
CN104239288A (en) * | 2014-09-18 | 2014-12-24 | 清华大学 | Method and system for checking and counseling psychological pressure of adolescents based on microblog |
CN104616666A (en) * | 2015-03-03 | 2015-05-13 | 广东小天才科技有限公司 | Method and device for improving dialogue communication effect based on speech analysis |
EP2874110A1 (en) * | 2013-11-15 | 2015-05-20 | Telefonica Digital España, S.L.U. | A method and a system to obtain data from voice analysis in a communication and computer programs products thereof |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101334196B1 (en) * | 2010-10-28 | 2013-11-28 | (주)아크릴 | Intelligent Affect Deducing Apparatus and Method therefor |
-
2015
- 2015-09-11 CN CN201510580511.XA patent/CN105206284B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101075435A (en) * | 2007-04-19 | 2007-11-21 | 深圳先进技术研究院 | Intelligent chatting system and its realizing method |
EP2874110A1 (en) * | 2013-11-15 | 2015-05-20 | Telefonica Digital España, S.L.U. | A method and a system to obtain data from voice analysis in a communication and computer programs products thereof |
CN104050256A (en) * | 2014-06-13 | 2014-09-17 | 西安蒜泥电子科技有限责任公司 | Initiative study-based questioning and answering method and questioning and answering system adopting initiative study-based questioning and answering method |
CN104239288A (en) * | 2014-09-18 | 2014-12-24 | 清华大学 | Method and system for checking and counseling psychological pressure of adolescents based on microblog |
CN104616666A (en) * | 2015-03-03 | 2015-05-13 | 广东小天才科技有限公司 | Method and device for improving dialogue communication effect based on speech analysis |
Non-Patent Citations (1)
Title |
---|
《Towards a Micro-Blog Platform for Sensing and Easing Adolescent Psychological Pressures》;Yuanyuan Xue et al.;《Poster, Demo, & Video Presentations》;20130912;第215-218页 |
Also Published As
Publication number | Publication date |
---|---|
CN105206284A (en) | 2015-12-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105206284B (en) | Dredge the cyberchat method and system of adolescent psychology pressure | |
CN108536681B (en) | Intelligent question-answering method, device, equipment and storage medium based on emotion analysis | |
Rivas et al. | Finding themes in qualitative data | |
Alamer et al. | Exploratory structural equation modeling in second language research: An applied example using the dualistic model of passion | |
Lai et al. | Natural language inference from multiple premises | |
CN109918650B (en) | Interview intelligent robot device capable of automatically generating interview draft and intelligent interview method | |
Visser et al. | Annotating argument schemes | |
EP3096246A1 (en) | Method, system and storage medium for realizing intelligent answering of questions | |
CN108829682B (en) | Computer readable storage medium, intelligent question answering method and intelligent question answering device | |
CN103970806B (en) | Method and device for establishing lyric emotion classification model | |
CN105912629A (en) | Intelligent question and answer method and device | |
Keraron et al. | Project PIAF: Building a native French question-answering dataset | |
Aparicio et al. | Emotion analysis of Portuguese political parties communication over the COVID-19 pandemic | |
CN114547293A (en) | Cross-platform false news detection method and system | |
Saiyed et al. | Technology-assisted motivational interviewing: developing a scalable framework for promoting engagement with tobacco cessation using NLP and machine learning | |
Fu et al. | When confidence and competence collide: Effects on online decision-making discussions | |
Dattner | Mapping Hebrew dative constructions | |
Amini et al. | Word associations contribute to machine learning in automatic scoring of degree of emotional tones in dream reports | |
Ehrlich et al. | Tracking the Evolution of Stare Decisis | |
CN106202052A (en) | Adolescent psychology pressure range and pressure source event perception method and device thereof | |
Rispoli | Cross-morpheme facilitation: The systematic emergence of agreement in 2-year-olds | |
Smeets et al. | Character centrality in present-day Dutch literary fiction | |
CN103810167B (en) | The method and apparatus for obtaining information | |
Caruso et al. | Do Associations Between Mind Wandering and Learning from Complex Texts Vary by Assessment Depth and Time? | |
Taylor et al. | Transformative diplomacy: negotiation of the WHO global code of practice on the international recruitment of health personnel |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |