CN103268311A - Event-structure-based Chinese statement analysis method - Google Patents
Event-structure-based Chinese statement analysis method Download PDFInfo
- Publication number
- CN103268311A CN103268311A CN2012104390074A CN201210439007A CN103268311A CN 103268311 A CN103268311 A CN 103268311A CN 2012104390074 A CN2012104390074 A CN 2012104390074A CN 201210439007 A CN201210439007 A CN 201210439007A CN 103268311 A CN103268311 A CN 103268311A
- Authority
- CN
- China
- Prior art keywords
- event
- verb
- statement
- role
- preposition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Machine Translation (AREA)
Abstract
The invention discloses an event-structure-based Chinese statement analysis method, which is used for expressing a Chinese statement into a tree structure consisting of a plurality of event designators and event characters comprising event character uncorrelated components corresponding to the event designators. The method comprises the following analysis steps of first preprocessing the Chinese statement; then performing event-structure-based Chinese statement analysis, finding the event designators in the statement and the event characters corresponding to the event designators, and analyzing multi-character components and the event character uncorrelated components; and finally labeling each statement component, adding function descriptions of each component, and giving a parenthesis representation form of a tree representation structure. The method is more applied to Chinese syntactic analysis, relationships between event elements are reflected, support is provided for semantic synthesis, and a language representation method and representation rules of an event can be reflected. According to the method, the event designators which are not verbs can be processed, and semantic functions of modifiers, prepositions, conjunctions and other components can be described.
Description
Technical field
The present invention relates to a kind of Chinese statement analytical method based on event structure, belong to natural language processing (Natural Language Processing) field.
Background technology
A large amount of existence are (little of an action to the description of all kinds of events among the human lives in the natural language, greatly to a historical events), also comprise simultaneously the description of content such as relation between role, state and the event of time that event produces, place, participation and feature.Chinese statement analytical method based on event structure of the present invention namely is each composition the statement is analyzed and to be shone upon from the angle of understanding event structure, and represents with formal method.Different language has different method for expressing and expression rule.People can be familiar with and understand the world by the relation between understanding event and the event.One of this important content that also to be natural language processing field pay close attention to about the research institute of event.Rise along with the internet correlation technique, people more and more depend on network and obtain information, and the information of internet presents characteristics such as magnanimity, sharp increase and redundancy, in order to monitor and use information wherein, allow machine can analyze event in the text, event-oriented statement analysis and research seem more and more important.(the Defense Advanced Research Projects Agency of the U.S. national defense Advanced Study Project council, DARPA) the topic recognition and tracking of sponsoring (Topic Detection and Tracking, TDT) evaluation and test meeting, its purpose is exactly to develop a series of information organizational technologies based on event.National Natural Science Foundation of China lists great project in 2012 in " unconventional accident contingency management research ".
The statement analysis just refers to each composition function and semanteme in the statement are analyzed, and with the linear word order between the word in the input sentence, becomes a nonlinear data structure.Statement analysis based on event structure just refers to each composition in the statement is mapped to described event structure or form.Analyzing according to different event and event role exactly specifically, is the tree structure that some event deictic words and corresponding with it event role comprise the relevant composition of non-event role with a Chinese sentence expression.Be the only way which must be passed of a nearly step realizing semantic understanding from the angle of understanding event to the statement analysis that a sentence carries out based on event structure.Natural language processing field comprises about the main theory of statement analysis at present: the formal grammar theory of interdependent sentence structure, Chomsky development is phrase structure grammar and expansion thereof, as: the phrase structure grammar that Lexical-Functional Grammar, functional unification grammar, Generalized Phrase Structure Grammar, centre word drive.The thought of these methods all is based upon on the English knowledge of grammar basis, the composition the sentence is not divided into event and event role and analyzes relation between them from the angle of understanding event.At present concentrate on from text identification and extraction event mostly for the research of event and the event role extracts, based on the automatic abstract of event and text aspect such as generations automatically, these study the support that all presses for the statement analytical method based on event structure of the present invention.That adopts that the present invention makes up not only can provide first hand characteristic information for machine learning and data mining based on the corpus of the Chinese statement analysis of event structure, can also be used for carrying out data statistics, make up statistics or probability model, extract language rule, the most important thing is that it will provide the standard of comparison and evaluation and test for the information processing technology based on event.
Though present existing statement analytical method has reflected the syntactic structure of sentence to a certain extent, but also there is not a kind of method from the angle of analyzing event the composition the sentence to be analyzed, and have the following disadvantages: (1) is not applicable to the statement analysis (Chinese is the language that meaning is closed) of Chinese entirely centered by grammatical function is analyzed; (2) centered by verb, can not cover all events (as: event noun " earthquake ", " fire " etc.) in the sentence; (3) lack semantic function explanation to ornamental equivalent, preposition, conjunction and other composition.
Summary of the invention
The problem and shortage that exists of prior art in view of the above, the present invention proposes the Chinese statement analytical method based on event structure, this method is on the basis of lexical analysis and the analysis of interdependent statement, from analysis event and event role's angle the composition the statement is divided into event deictic words, event role, ornamental equivalent, preposition, conjunction and other word composition that the event deictic words are corresponding again, by the label setting, and add the semantic function explanation of each composition, sentence structure is converted to tree construction based on event and role thereof by linear order.
Chinese statement analytical method based on event structure of the present invention is achieved through the following technical solutions, and specifically comprises:
A, statement pre-treatment step: adopt morphology and grammatical analysis instrument statement to be carried out the pre-service of participle, part-of-speech tagging, interdependent syntactic analysis;
B, based on the Chinese statement analytical procedure of event: pretreated statement is carried out Chinese statement analysis based on event;
C, interpolation label and function declaration step: the statement after analyzing is added label and function declaration, and the mark object comprises event deictic words, event role and three principal ingredients of non-event role.
The described statement pre-service of above-mentioned steps A adopts the social statement of language technology platform LTP that provides with the Research into information retrieval center that calculates of Harbin Institute of Technology to carry out participle, part-of-speech tagging (adopting China national 863 evaluation and test part of speech label sets), interdependent syntactic analysis processing.Statement after the machine processing is that unit mark has word number, part of speech, dependence information with the word.
The described Chinese statement analysis based on event structure of above-mentioned steps B, pretreated statement is carried out Chinese statement analysis based on event structure, comprise: determine in the statement the event deictic words, determine each event deictic words the event key element, determine modification event deictic words and event role ornamental equivalent, determine many identity sentence element, determine other sentence element that concrete steps are as follows:
B1, determine the event deictic words step in the statement: find out all verbs and event proper noun in the sentence, verb is classified, verb type comprises sincere verb and abstract verb, analyze the dependence of each verb, if the dependence of verb is for concerning in fixed that ATT or several verb are that the dependence of one of them verb of coordination COO is ATT then these verbs are ornamental equivalent, remaining event proper noun and verb all are regarded as the event deictic words.
B2, determine event role's step of each event deictic words: find the event role of corresponding each the event deictic words person that is the agent, word denoting the receiver of an action person, environment, time, instrument etc. by analysis, and the part of serving as the identical or different event role of a plurality of event deictic words in the sentence is found out.
B3, determine modification event deictic words and event role's ornamental equivalent step: find out by semantic analysis and dependency analysis and have the part of modifying implication, and determine polygonal color part again.
B4, determine other sentence element step: by remaining composition in the sentence of above analysis back preposition, conjunction, idiom, interjection, onomatopoeia, morpheme, non-lexeme and punctuation mark are arranged, choose preposition and conjunction and carry out semantic analysis, other composition wouldn't process, and changing rule is other composition.
The described interpolation label of above-mentioned steps C and function declaration, statement after analyzing through step B is added label and semantic function explanation, comprise: the polygonal look sentence element of mark, flag event role's composition, the non-event role's composition of mark, based on the formalization representation of the Chinese statement parsing tree of event structure, concrete steps are as follows:
C1, the polygonal look sentence element of mark step: according to the analysis result of step B, at first polygonal look sentence element mark is come out, the phase label is " MC " (Multiple Character), adds polygonal look sentence element numbering " mcID " then, and wherein ID is a natural number.
C2, flag event role become step by step: according to the analysis result of step B, and each event role composition and numbering in the polygonal colour content of mark one by one, and then mark other event role; As not having polygonal colour content then direct mark each event role composition and label in the sentence.Event role comprises: event deictic words label " denote ", numbering " eID ", agent person label " subject ", numbering " sID ", word denoting the receiver of an action person label " object ", numbering " oID ", environmental labels " loctor ", numbering " lID ", time tag " time ", numbering " tID ", instrument label " tool ", numbering " toID ".
C3, the non-event role of mark become step by step: according to the analysis result of step B, label and the function declaration of other sentence element in the whole sentence of mark and the polygonal colour content comprise: ornamental equivalent label " modifier ", preposition label " preposition ", conjunction label " conjunction ", other composition label " others ".
C4, based on the formalization representation step of the Chinese statement parsing tree of event structure: by the processing of C1, C2, C3 step, whole sentence is described each event role and non-event role composition with the angle of analyzing event with tree structure, will show based on the Chinese statement parsing tree of event structure by parenthesized notation at last.
Chinese statement analytical method based on event structure of the present invention is compared the advantage that has with traditional statement analytical method and is: (1) this method is applicable to the statement analysis of Chinese, and the relation between the reflection event role is for semantic analysis provides support; (2) this method can reflect language representation's method and the expression rule of event; (3) this method can be handled the event deictic words of non-verb; (4) this method describes the function of ornamental equivalent, preposition, conjunction and other composition.
Description of drawings
Fig. 1 is the process flow diagram of the Chinese statement analytical method based on event structure of the present invention;
Fig. 2 is the described process flow diagram of analyzing based on the Chinese statement of event structure of step B among Fig. 1;
Fig. 3 is the process flow diagram of the described interpolation label of step C and function declaration among Fig. 1.
Embodiment
The present invention is described in further detail with preferred enforcement below in conjunction with accompanying drawing.
As shown in Figure 1, this is based on the Chinese statement analytical method of event structure, and it specifically may further comprise the steps:
A, statement pre-treatment step, machine participle, part-of-speech tagging, interdependent syntactic analysis: adopt society of Harbin Institute of Technology that the statement of language technology platform LTP2.1 that provides with the Research into information retrieval center is provided and carry out participle, part-of-speech tagging---adopt China national 863 evaluation and tests to handle with part of speech label sets, interdependent syntactic analysis, statement mark statement number after the machine processing, word number, with tense marker part of speech and the dependence of each word.For example, the effect after a statement is handled through participle, part-of-speech tagging and interdependent syntactic analysis is as follows:
Example 1: Xiao Wang believes that Xiao Zhang will come.
Pretreated effect following (word is that word label, id record word number, cont record word character string, pos record part of speech mark, parent record father node id number, relate records dependence, and wherein nh represents that name, v represent that verb, d represent that adverbial word, wp represent that punctuation mark, SBV represent that subject-predicate relation, HED represent that head verb, VOB represent that moving guest concerns, ADV represents to concern in the shape):
" Xiao Wang believes that Xiao Zhang will come to＜sent id=" 0 " cont=.">
＜word id=" 0 " cont=" Xiao Wang " pos=" nh " parent=" 1 " relate=" SBV "/〉
＜word id=" 1 " cont=" believe " pos=" v " parent=" 1 " relate=" HED "/
＜word id=" 2 " cont=" Xiao Zhang " pos=" nh " parent=" 1 " relate=" VOB "/〉
＜word id=" 3 " cont=" with " pos=" d " parent=" 4 " relate=" ADV "/〉
＜word id=" 4 " cont=" come " pos=" v " parent=" 1 " relate=" VOB "/
<word?id="5"?cont="。"?pos="wp"?parent="-2"?relate="WP"?/>
</sent>
B, based on the Chinese statement analytical procedure of event structure, with reference to Fig. 2: pretreated statement is carried out Chinese statement analysis based on event structure, and concrete steps are as follows;
B1, determine the event deictic words in the statement: find out all verbs and event proper noun in the sentence, verb is classified, verb type comprises sincere verb and abstract verb, analyze the dependence of each verb, if the dependence of verb is for concerning in fixed that ATT or several verb are that the dependence of one of them verb of coordination COO is ATT then these verbs are ornamental equivalent, remaining event proper noun and verb are the event deictic words, specify respectively below by the example of determining event deictic words in the statement:
B11, event proper noun
The event proper noun is the special noun of a class, has represented the generation of certain event in statement, for example, below have literal " earthquake " and " fire " of underscore in the example.
14: 28 on the 12nd 2:5 month of example, the Wenchuan, Sichuan takes place 7.8 grades
Earthquake
Example 3: her appearance exists
FireIn seriously damaged.
B12, sincere verb
Sincere verb is general verb, is used for representing action or behavior itself, possesses the main grammar property of verb, is typical verb, for example has literal " swimming " and " study " of underscore in the following sentence.
Example 4: he is shouting to say to us and is not going to river
Swimming
Example 5: Xiao Ming
StudyChinese language.
The B13 abstract verb
Abstract verb is the verb of other type outside the sincere verb, for example, below in the example band underscore the literal "Yes" and " should ".
Example 6: this this book
BeI.
Example 7: we
ShouldFrom others' mistake, learn the lesson.
The B14 dependence concerns that the verb of ATT plays the verb of modification in fixed
Dependence is that the verb of ATT represents that this verb does ornamental equivalent in sentence, does not consider as the event deictic words.Same several verb is that the dependence of one of them verb of coordination COO is ATT, represents that these several verbs are all ornamental equivalent.For example, literal " relevant ", " cooling " and " dilution " of band underscore in the example below.
Example 8: principal leads
RelevantThe responsible official of institute participates in recreational activities.
Example 9: firefighters passes through
Cooling,
DilutionEtc. the method emergency treatment.
B2, determine event role's step of each event deictic words: find the event role of corresponding each the event deictic words person that is the agent, word denoting the receiver of an action person, environment, time, instrument by analysis, and the similar and different event key element of serving as a plurality of event deictic words in the sentence is partly found out.Specify respectively below by the event role instance of determining event deictic words correspondence in the statement.
B21, event role's agent person
The main body that the agent person namely moves, expression applies the people of action or thing etc., for example, below have underscore in the example literal " they " be the main body of event deictic words " dancing ".
Example 10:
TheyDance in a round.
B22, event role's word denoting the receiver of an action person
The object that the word denoting the receiver of an action person namely moves, expression are moved people or the thing etc. of domination, for example, below in the example literal " doggie " of band underscore be the object that the event deictic words " take in one's arms ".
Example 11: he takes in one's arms
DoggieAnd it has been taken to safe place.
B23, event role environment
Environment is namely described information such as place that action takes place, position, for example, below in the example literal " Wenchuan, Sichuan " of band underscore be the place of event deictic words " generation " and " earthquake ".
14: 28 on the 12nd 12:5 month of example,
The Wenchuan, Sichuan7.8 grades of earthquakes take place.
B24, event role's time
Time is namely described the time that action takes place, can be absolute time, relative time or time interval, for example, below in the example literal " 14: 28 on the 12nd May ", " tomorrow morning " and " two weeks " of band underscore be respectively the time of event deictic words " generation " and " earthquake ", " going ", " going on business ".
Example 13:
14: 28 on the 12nd May, 7.8 grades of earthquakes take place in the Wenchuan, Sichuan.(absolute time)
Example 14: I
Tomorrow morningGo to school.(relative time)
Example 15: Mr. Li will go on business
Two weeks(time interval)
B25, event role's instrument
Instrument namely moves the instrument that adopts, for example, below in the example literal " chalk " of band underscore be the instrument that the event deictic words " write " and use.
Example 16: teacher uses
ChalkWrite at blackboard.
B26, many identity event role
Many identity event role is that a certain partial content in the sentence serves as the corresponding event role of different event deictic words simultaneously, for example, the literal " tent " of band underscore is the word denoting the receiver of an action person of Event triggered word " allocation and transportation " in the following example, also is the word denoting the receiver of an action person that the Event triggered word " is delivered to ".
Example 17: urgent allocation and transportation
TentDelivered to the disaster area.
B3, determine modification event deictic words and event role's ornamental equivalent step: find out the part with modification implication by semantic analysis and dependency analysis, generally mainly investigate dependence and concern ATT in fixed, quantitative relation QUN, voice structure MT, " " word structure DE, " " word structure DI, the word of verbal endocentric phrase ADV, and determine whether comprise the event role who had analyzed in the ornamental equivalent again, if comprise, then this ornamental equivalent is polygonal color part, for example, comprise event deictic words " allocation and transportation " in the literal of band underscore " urgent allocation and transportation " in the following example, also having " divorce " this word namely to modify " woman " this word, is again event deictic words.
Example 18:
Urgent allocation and transportationTent has been delivered to the disaster area.
Example 19:
DivorceThe woman even deceives muddled man.
B4, determine other sentence element step: by remaining composition in the sentence of above analysis back preposition, conjunction, idiom, interjection, onomatopoeia, morpheme, non-lexeme and punctuation mark are arranged.Other composition wouldn't process except preposition and conjunction.Functional analysis to preposition and conjunction is namely classified to it.Preposition be divided into expression time (as: " since "), place (as: " "), direction (as: " to "), mode (as: " by "), method (as: " according to "), according to (as: " according to "), instrument (as: " with "), relatively (as: " ratio "), reason (as: " because of "), purpose (as: " for "), agent (as: " "), word denoting the receiver of an action (as: " quilt "), concern object (as: " for ") and other totally 14 types, conjunction is divided into.Conjunction is divided into side by side (as: " with "), accepts (as: " then "), turnover (as: " still "), cause and effect (as: " therefore "), select (as: " perhaps "), suppose (as: " if "), (as: " being not so good as "), concession (as: " immediately "), go forward one by one (as: " not only "), condition (as: " needing only "), purpose (as: " so that ") and other conjunction in totally 12 relatively.
C, add label and semantic analysis description of step, with reference to Fig. 3, the statement after analyzing is carried out mark and function declaration, distinguish flag event role composition, the non-event role's composition of mark, based on the formalization representation of the statement parsing tree of event structure.Tag format underlined in the present embodiment all adopts the XML language, and concrete steps are as follows:
C1, the polygonal look sentence element of mark step: according to the analysis result of step B, at first polygonal look sentence element mark is come out, its label is " MC " (Multiple Character), add polygonal look sentence element numbering " mcID " then, wherein ID is a natural number, for example, the literal " Xiao Zhang will come " of band underscore namely is the object that the event deictic words " are believed " in the following example, " Xiao Zhang " wherein is again the main body that the event deictic words " are come ", so with "＜MC〉" beginning, the MC of "＜/〉 " finish it is carried out mark, specifically be expressed as follows.
Example 20: Xiao Wang believes
Xiao Zhang will come
" Xiao Wang believes that Xiao Zhang will come to＜sent id=" 0 " cont=.">
＜word id=" 0 " cont=" Xiao Wang " pos=" nh " parent=" 1 " relate=" SBV "/〉
＜word id=" 1 " cont=" believe " pos=" v " parent=" 1 " relate=" HED "/
<MC?mcid="mc1">
＜word id=" 2 " cont=" Xiao Zhang " pos=" nh " parent=" 1 " relate=" VOB "/〉
＜word id=" 3 " cont=" with " pos=" d " parent=" 4 " relate=" ADV "/〉
＜word id=" 4 " cont=" come " pos=" v " parent=" 1 " relate=" VOB "/
<word?id="5"?cont="。"?pos="wp"?parent="-2"?relate="WP"?/>
</MC>
</sent>
C2, flag event role become step by step: according to the analysis result of step B, and each event role composition and the numbering in the polygonal colour content of mark, mark other event role and numbering then one by one; As not having polygonal colour content then direct mark each event role composition and numbering in the sentence.In the numbering of event deictic words " eID " ID number determined by the type of event deictic words and the degree of depth in dependency tree, its priority rule is: the event proper noun〉verb 1 is 1 layer of dependency tree〉verb 2 is 2 layers of dependency trees ... verb n is interdependent book leaf node, one deck if several verb coexists then arrange according to order from left to right.Except the event deictic words, other event role's numbering is determined by the numbering of its corresponding event deictic words, except flag event role's label and numbering, also wants some function declarations of mark, and concrete tag content is described as follows:
C21,＜subject(agent person label) the sid(numbering)=" sID " t_subject(type)=" creature(people or biology) | the things(thing) | organization(organizational structure) | the phrase(phrase) | the clause(short sentence) | the event(event) "＜/subject〉(end mark)
C22,＜object(word denoting the receiver of an action person label) the oid(numbering)=" oID " t_object(type)=" creature(people or biology) | the things(thing) | organization(organizational structure) | the phrase(phrase) | the clause(short sentence) | the event(event) "＜/object〉(end mark)
C23, <denote (event indicates that the word label), Eid (number) = "eID", T_denote (type) = "event_v (real meaning verb) | sense_v (abstract verb) | event_n (event proper noun)", Tendency (verb type tendencies) = ", VX (judgment verb) | VM (psychological verbs), | VD (verbs) | VO (Modal Verb) | VF (Imperative verb) | VV (confession verb) | VA (verb behavior ) | VM (than you compare verb) | VE (usually a verb) |, proprietary, (event proprietary word) ", performance (action completion) =" happen (has happened) | unhappen (no) | happing (being happen) ", wordtime (action event) =" bygone (past) | now (now) | future future ">, </ denote> (end marker)
C24,＜the time(time tag) the tid(numbering)=" tID " t_time(type)=" the absTime(absolute time) | the relTime(relative time) | the timeInterval(time interval) "〉＜/time〉(end mark)
C25,＜the locotr(environmental labels) the lid(numbering)=" lID " t_loctor(type)=" the origin(departure place) | the destination(destination) | place(environment place) "〉＜/loctor〉(end mark)
C26,＜tool(instrument label) the toid(numbering)=" tID " t_tool(tool types)=" creature(people or biology) | the thing(thing) | the event(event) "〉＜/tool〉(end mark)
Specify the concrete enforcement of C2 step below by example:
Example 21: 1 pretreated result carries out mark to example, and the result is as follows:
" Xiao Wang believes that Xiao Zhang will come to＜sent id=" 0 " cont=.">
<subject?sid="s1"?t_subject="creature?">
＜word id=" 0 " cont=" Xiao Wang " pos=" nh " parent=" 1 " relate=" SBV "/〉
</subject>
<denote?eid="e1"?t_denote="sense_v"?tendency="VP"?performance="happen"?wordtime="bygone">
＜word id=" 1 " cont=" believe " pos=" v " parent=" 1 " relate=" HED "/
</denote>
<MC?mcid="mc1"?>
<object?oid="o1"?t_object="clause">
<subject?sid="s2"?t_subject="creature">
＜word id=" 2 " cont=" Xiao Zhang " pos=" nh " parent=" 1 " relate=" VOB "/〉
</subject>
＜word id=" 3 " cont=" with " pos=" d " parent=" 4 " relate=" ADV "/〉 (ornamental equivalent will be handled in the C3 step)
<denote?eid="e2"?t_denote="event_v"?tendency="VD"?performance="unhappen?"?wordtime="future">
＜word id=" 4 " cont=pos=" v " parent=" 1 " relate=" VOB " that " comes " 〉
</denote>
</object>
</MC>
<word?id="5"?cont="。"?pos="wp"?parent="-2"?relate="WP"?/>
</sent>
C3, the non-event role of mark become step by step: according to the analysis result of step B, and label and the function declaration of the non-event role's composition in the whole sentence of mark, concrete tag content is described as follows:
C31, <modifier (modifier ingredient labels), M_element (modified ingredients) = "eID (event directive number) | sID (causal agent ID) | oID (Patient with ID) | tID (event number) | lID (environmental Code) | toID (tool number) | mcID (multi-role component ID) ", t_modifier (modifier type) =" adjective (adjective) | adverb (adverb) | phrase (phrase) | clause (phrase) | noun (noun) | verb (verb) | proprietary (exclusive events) |, auxiliary (particle), | others (other) ", m_appraise (modified evaluation direction) =" commendatory (praise) | pejorative (derogatory) | neutral (in Xing) | bygone (past) | now (now) | future (future) | degree (degree) | quality (quality) | quantity (amount) | time (hours) | speed (velocity) | affiliation (membership) |, tense (tense), | negative (negative) |, frequentness (frequency) | post (job) | pattern (mode) | method (method) ......, ">, </ modifier> (end marker)
C32, <conjuction (conjunctions label), Cid (number) = "cID", S_conjunction (even the word order) = "beg (starting conjunctions) | mid (middle conjunctions) | end (end of conjunctions) |, Single (single conjunctions ) ", t_conjunction (type), =" coordinating (parallel relationship) | continue (to undertake relations) | transition (transition relation) | karma (causal) | select (select relations) | suppose (assuming relations) | compare (comparison between ) | concession (concession relations) | progressive (progressive relationship) | conditional (condition relations) | purpose (purpose relations) "> </ conjuction> (end marker)
C33,＜preposition (preposition label) t_prepositon(preposition type)=" time_p(time preposition) | loctor_p(place preposition) | pattern_p(mode preposition) | method_p (method preposition) | accord_p(is according to preposition) | tool_p(instrument preposition) | compare_p(is preposition relatively) | reason_p(reason preposition) | objective_p(purpose preposition) | subject_p(agent preposition) | object_p(word denoting the receiver of an action preposition) | involve_p(concerns the object preposition) "〉＜/preposition〉(end mark)
C34,＜others (other composition label) t_others(type)=" the idiom(idiom) | the exelamation(interjection) | the onomatopoetic(onomatopoeia) | the morpheme(morpheme) | the non-lexeme of non-lexeme() | the prefix(prefix) | the suffix(suffix) | DE(" " structure) | DI(" " structure) | ... " the others of〉＜/〉 (end mark)
Specify the concrete enforcement of C3 step below by example:
Example 22: the mark result to example 21 continues to handle as follows
" Xiao Wang believes that Xiao Zhang will come to＜sent id=" 0 " cont=.">
<subject?sid="s1"?t_subject="creature?">
＜word id=" 0 " cont=" Xiao Wang " pos=" nh " parent=" 1 " relate=" SBV "/〉
</subject>
<denote?eid="e1"?t_denote="sense_v"?tendency="VP"?performance="happen"?wordtime="bygone">
＜word id=" 1 " cont=" believe " pos=" v " parent=" 1 " relate=" HED "/
</denote>
<MC?mcid="mc1"?>
<object?oid="o1"?t_object="clause">
<subject?sid="s2"?t_subject="creature">
＜word id=" 2 " cont=" Xiao Zhang " pos=" nh " parent=" 1 " relate=" VOB "/〉
</subject>
<modifier?m_element="e2"?t_modifier="?adverb"?m_appraise="tense">
＜word id=" 3 " cont=" with " pos=" d " parent=" 4 " relate=" ADV "/〉
</modifier>
<denote?eid="e2"?t_denote="event_v"?tendency="VD"?performance="unhappen?"?wordtime="future">
＜word id=" 4 " cont=pos=" v " parent=" 1 " relate=" VOB " that " comes " 〉
</denote>
</object>
</MC>
<word?id="5"?cont="。"?pos="wp"?parent="-2"?relate="WP"?/>
</sent>
C4, based on the formalization representation step of the statement parsing tree of event structure: by the processing of C1, C2, C3 step, whole sentence is described each event role and non-event role composition with the angle of analyzing event with tree structure, will show based on the statement parsing tree of event structure by parenthesized notation at last.For example, the tree construction with example 22 shows with parenthesized notation below, and is specific as follows:
Example 23:root(s1-e1-mc1 (o1 (s2-m (e2)-e2)))
Show more detailed information if desired, can also record each event role's function declaration with square bracket " [] ", be expressed as follows:
root（s1[creature]-e1[sense_v-VP-happen-bygone]-
mc1?(o1(s2[creature]-m[adverb-tense]?(e2)-
e2[event_v-VD-unhappen-future]))）
More than the Chinese statement analytical method based on event structure of the present invention has been done detailed explanation.Modification and improvement that those skilled in the art make in design scope of the present invention should be included in the appended claim restricted portion of the present invention.
Claims (4)
1. based on the Chinese statement analytical method of event structure, it is characterized in that: with a Chinese sentence expression be some event deictic words and with it corresponding event role comprise the tree structure of the relevant composition of non-event, the Chinese statement analysis concrete operations step of event structure is as follows:
A, statement pre-service: adopt morphology and grammatical analysis instrument statement to be carried out the pre-service of participle, part-of-speech tagging, interdependent syntactic analysis;
B, based on the Chinese statement analysis of event: pretreated statement is carried out Chinese statement analysis based on event; Find out event deictic words and corresponding event role with it in the statement, analyze the relevant composition with non-event of polygonal colour content;
C, interpolation label and function declaration: the statement after analyzing is added label and function declaration, and the mark object comprises event deictic words, event role and three principal ingredients of non-event role, provides the bracketed form of tree represenation structure.
2. the Chinese statement analytical method based on event structure according to claim 1, it is characterized in that: the pretreated concrete operation method of the described statement of described steps A is as follows:
Adopt society of Harbin Institute of Technology that the statement of language technology platform LTP2.1 that provides with the Research into information retrieval center is provided and carry out participle, part-of-speech tagging---adopt China national 863 evaluation and tests to handle with part of speech label sets, interdependent syntactic analysis, statement mark statement number after the machine processing, word number, with tense marker part of speech and the dependence of each word.
3. the Chinese statement analytical method based on event structure according to claim 1 is characterized in that, the described concrete implementation step of analyzing based on the Chinese statement of event of described step B is as follows:
B1, determine the event deictic words in the statement: find out all verbs and event proper noun in the sentence, verb is classified, verb type comprises sincere verb and abstract verb, analyze the dependence of each verb, if the dependence of verb is that the dependence of one of them verb of coordination (COO) is ATT then these verbs are ornamental equivalent for relation (ATT) or several verb in fixed, remaining event proper noun and verb are the event deictic words, below to determining the explanation of event deictic words in the statement:
B11, event proper noun
The event proper noun is the special noun of a class, has represented the generation of certain event in statement;
B12, sincere verb
Sincere verb is general verb, is used for representing action or behavior itself, possesses the main grammar property of verb, is typical verb;
The B13 abstract verb
Abstract verb is the verb of other type outside the sincere verb;
The B14 dependence is the verb that ATT plays modification
Dependence is represented this verb for the verb that concerns ATT in fixed and do ornamental equivalent in sentence, do not consider as the event deictic words; Same several verb is that the dependence of one of them verb of coordination COO is ATT, represents that these several verbs are all ornamental equivalent;
B2, determine the event role of each event deictic words: find the event role of corresponding each the event deictic words person that is the agent, word denoting the receiver of an action person, environment, time, instrument by analysis, and the similar and different event key element of serving as a plurality of event deictic words in the sentence is partly found out;
B21, event role's agent person
The main body that the agent person namely moves, expression applies people or the thing of action;
B22, event role's word denoting the receiver of an action person
The object that the word denoting the receiver of an action person namely moves, the expression moved the domination people or thing;
B23, event role environment
Environment is namely described information such as place that action takes place, position;
B24, event role's time
Time is namely described the time that action takes place, and can be absolute time, relative time or time interval;
B25, event role's instrument
Instrument namely moves the instrument that adopts;
B26, many identity event role
Many identity event role is that a certain partial content in the sentence serves as the corresponding event role of different event deictic words simultaneously;
B3, determine modification event deictic words and event role's ornamental equivalent: find out by semantic analysis and dependency analysis and have the part of modifying implication, generally mainly investigate dependence for concern in fixed ATT, quantitative relation QUN, voice structure MT, " " word structure DE, " " word of word structure DI, verbal endocentric phrase ADV, and determine whether comprise the event role who had analyzed in the ornamental equivalent again, if comprise, then this ornamental equivalent is polygonal color part;
B4, determine other sentence element: by remaining composition in the sentence of above analysis back preposition, conjunction, idiom, interjection, onomatopoeia, morpheme, non-lexeme and punctuation mark are arranged; Other composition wouldn't process except preposition and conjunction; Functional analysis to preposition and conjunction is namely classified to it; Preposition is divided into expression time, place, direction, mode, method, foundation, instrument, comparison, reason, purpose, agent, word denoting the receiver of an action, concern object and other totally 14 types; Conjunction is divided into side by side, accepts, turnover, cause and effect, selection, suppose, comparison, give way, go forward one by one, condition, purpose and other conjunction in totally 12.
4. the Chinese statement analytical method based on event structure according to claim 1 is characterized in that, interpolation label and the function declaration described in the described step C, and underlined tag format all adopts the XML language, and its concrete implementation step is as follows:
C1, the polygonal look sentence element of mark: according to the analysis result of step B, at first polygonal look sentence element mark is come out, its label is " MC ", adds polygonal look sentence element numbering " mcID " then, and wherein ID is a natural number;
C2, flag event role's composition: according to the analysis result of step B, each event role composition and the numbering in the polygonal colour content of mark, mark other event role and numbering then one by one; As not having polygonal colour content then direct mark each event role composition and numbering in the sentence;
In the numbering of event deictic words " eID " ID number determined by the type of event deictic words and the degree of depth in dependency tree, its priority rule is: the event proper noun〉verb 1 is 1 layer of dependency tree〉verb 2 is 2 layers of dependency trees ... verb n is interdependent book leaf node, one deck if several verb coexists then arrange according to order from left to right; Except the event deictic words, other event role's numbering is determined by the numbering of its corresponding event deictic words, except flag event role's label and numbering, also wants some function declarations of mark, and concrete tag content is described as follows:
C21,＜subject(agent person label) the sid(numbering)=" sID " t_subject(type)=" creature(people or biology) | the things(thing) | organization(organizational structure) | the phrase(phrase) | the clause(short sentence) | the event(event) "＜/subject〉(end mark)
C22,＜object(word denoting the receiver of an action person label) the oid(numbering)=" oID " t_object(type)=" creature(people or biology) | the things(thing) | organization(organizational structure) | the phrase(phrase) | the clause(short sentence) | the event(event) "＜/object〉(end mark)
C23, <denote (event indicates that the word label), Eid (number) = "eID", T_denote (type) = "event_v (real meaning verb) | sense_v (abstract verb) | event_n (event proper noun)", Tendency (verb type tendencies) = ", VX (judgment verb) | VM (psychological verbs), | VD (verbs) | VO (Modal Verb) | VF (Imperative verb) | VV (confession verb) | VA (verb behavior ) | VM (than you compare verb) | VE (usually a verb) |, proprietary, (event proprietary word) ", performance (action completion) =" happen (has happened) | unhappen (no) | happing (being happen) ", wordtime (action event) =" bygone (past) | now (now) | future future ">, </ denote> (end marker)
C24,＜the time(time tag) the tid(numbering)=" tID " t_time(type)=" the absTime(absolute time) | the relTime(relative time) | the timeInterval(time interval) "〉＜/time〉(end mark)
C25,＜the locotr(environmental labels) the lid(numbering)=" lID " t_loctor(type)=" the origin(departure place) | the destination(destination) | place(environment place) "〉＜/loctor〉(end mark)
C26,＜tool(instrument label) the toid(numbering)=" tID " t_tool(tool types)=" creature(people or biology) | the thing(thing) | the event(event) "〉＜/tool〉(end mark)
C3, the non-event role's composition of mark: according to the analysis result of step B, label and the function declaration of the non-event role's composition in the whole sentence of mark, concrete tag content is described as follows:
C31, <modifier (modifier ingredient labels), M_element (modified ingredients) = "eID (event directive number) | sID (causal agent ID) | oID (Patient with ID) | tID (event number) | lID (environmental Code) | toID (tool number) | mcID (multi-role component ID) ", t_modifier (modifier type) =" adjective (adjective) | adverb (adverb) | phrase (phrase) | clause (phrase) | noun (noun) | verb (verb) | proprietary (exclusive events) |, auxiliary (particle), | others (other) ", m_appraise (modified evaluation direction) =" commendatory (praise) | pejorative (derogatory) | neutral (in Xing) | bygone (past) | now (now) | future (future) | degree (degree) | quality (quality) | quantity (amount) | time (hours) | speed (velocity) | affiliation (membership) |, tense (tense), | negative (negative) |, frequentness (frequency) | post (job) | pattern (mode) | method (method) ......, ">, </ modifier> (end marker)
C32, <conjuction (conjunctions label), Cid (number) = "cID", S_conjunction (even the word order) = "beg (starting conjunctions) | mid (middle conjunctions) | end (end of conjunctions) |, Single (single conjunctions ) ", t_conjunction (type), =" coordinating (parallel relationship) | continue (to undertake relations) | transition (transition relation) | karma (causal) | select (select relations) | suppose (assuming relations) | compare (comparison between ) | concession (concession relations) | progressive (progressive relationship) | conditional (condition relations) | purpose (purpose relations) "> </ conjuction> (end marker)
C33,＜preposition (preposition label) t_prepositon(preposition type)=" time_p(time preposition) | loctor_p(place preposition) | pattern_p(mode preposition) | method_p (method preposition) | accord_p(is according to preposition) | tool_p(instrument preposition) | compare_p(is preposition relatively) | reason_p(reason preposition) | objective_p(purpose preposition) | subject_p(agent preposition) | object_p(word denoting the receiver of an action preposition) | involve_p(concerns the object preposition) "〉＜/preposition〉(end mark)
C34,＜others (other composition label) t_others(type)=" the idiom(idiom) | the exelamation(interjection) | the onomatopoetic(onomatopoeia) | the morpheme(morpheme) | the non-lexeme of non-lexeme() | the prefix(prefix) | the suffix(suffix) | DE(" " structure) | DI(" " structure) | ... " the others of〉＜/〉 (end mark)
C4, based on the formalization representation of the statement parsing tree of event structure: by the processing of C1, C2, C3 step, whole sentence is described each event role and non-event role composition with the angle of analyzing event with tree structure, will show based on the statement parsing tree of event structure by parenthesized notation at last.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012104390074A CN103268311A (en) | 2012-11-07 | 2012-11-07 | Event-structure-based Chinese statement analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012104390074A CN103268311A (en) | 2012-11-07 | 2012-11-07 | Event-structure-based Chinese statement analysis method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103268311A true CN103268311A (en) | 2013-08-28 |
Family
ID=49011942
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012104390074A Pending CN103268311A (en) | 2012-11-07 | 2012-11-07 | Event-structure-based Chinese statement analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103268311A (en) |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104090867A (en) * | 2014-07-17 | 2014-10-08 | 北京中电拓方科技发展有限公司 | Method for executing event based on coal mine safety quality standard |
| CN105224640A (en) * | 2015-09-25 | 2016-01-06 | 杭州朗和科技有限公司 | A kind of method and apparatus extracting viewpoint |
| CN105243056A (en) * | 2015-09-07 | 2016-01-13 | 饶志刚 | Punctuation mark processing based Chinese syntax analysis method and apparatus |
| CN106021234A (en) * | 2016-05-31 | 2016-10-12 | 徐子涵 | Label extraction method and system |
| CN106897364A (en) * | 2017-01-12 | 2017-06-27 | 上海大学 | Chinese based on event refers to building of corpus method |
| CN107451158A (en) * | 2016-06-01 | 2017-12-08 | 中国科学院地理科学与资源研究所 | Traffic events semantic role abstracting method in a kind of network text |
| CN107679035A (en) * | 2017-10-11 | 2018-02-09 | 石河子大学 | A kind of information intent detection method, device, equipment and storage medium |
| CN108197294A (en) * | 2018-01-22 | 2018-06-22 | 桂林电子科技大学 | A kind of text automatic generation method based on deep learning |
| CN108595421A (en) * | 2018-04-13 | 2018-09-28 | 北京神州泰岳软件股份有限公司 | A kind of abstracting method, the apparatus and system of Chinese entity associated relationship |
| CN108959464A (en) * | 2018-06-19 | 2018-12-07 | 李勤骞 | Learning method and system containing auxiliary word |
| CN109471936A (en) * | 2018-10-11 | 2019-03-15 | 上海叔本华智能科技有限公司 | A kind of method and system for plant maintenance information progress tagsort |
| CN109815481A (en) * | 2018-12-17 | 2019-05-28 | 北京百度网讯科技有限公司 | Method, apparatus, equipment and the computer storage medium of event extraction are carried out to text |
| CN110781369A (en) * | 2018-07-11 | 2020-02-11 | 天津大学 | Emotional cause mining method based on dependency syntax and generalized causal network |
| CN110851560A (en) * | 2018-07-27 | 2020-02-28 | 杭州海康威视数字技术股份有限公司 | Information retrieval method, device and equipment |
| CN110874531A (en) * | 2020-01-20 | 2020-03-10 | 湖南蚁坊软件股份有限公司 | Topic analysis method and device and storage medium |
| CN111581954A (en) * | 2020-05-15 | 2020-08-25 | 中国人民解放军国防科技大学 | Text event extraction method and device based on grammar dependency information |
| CN112148838A (en) * | 2020-09-23 | 2020-12-29 | 北京中电普华信息技术有限公司 | Business source object extraction method and device |
| CN112699664A (en) * | 2021-01-08 | 2021-04-23 | 中国专利信息中心 | Chinese syntax analysis method and system |
| CN115017913A (en) * | 2022-04-21 | 2022-09-06 | 广州世纪华轲科技有限公司 | Semantic component analysis method based on master-slave framework mode |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101782897A (en) * | 2010-03-17 | 2010-07-21 | 上海大学 | Chinese corpus labeling method based on events |
| CN101937430A (en) * | 2010-09-03 | 2011-01-05 | 清华大学 | Method for extracting event sentence pattern from Chinese sentence |
| CN101957812A (en) * | 2010-09-21 | 2011-01-26 | 上海大学 | Verb semantic information extracting method based on event ontology |
| US20110257963A1 (en) * | 2006-10-10 | 2011-10-20 | Konstantin Zuev | Method and system for semantic searching |
| US20110270607A1 (en) * | 2006-10-10 | 2011-11-03 | Konstantin Zuev | Method and system for semantic searching of natural language texts |
| US20120010872A1 (en) * | 2006-10-10 | 2012-01-12 | Abbyy Software Ltd | Method and System for Semantic Searching |
-
2012
- 2012-11-07 CN CN2012104390074A patent/CN103268311A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110257963A1 (en) * | 2006-10-10 | 2011-10-20 | Konstantin Zuev | Method and system for semantic searching |
| US20110270607A1 (en) * | 2006-10-10 | 2011-11-03 | Konstantin Zuev | Method and system for semantic searching of natural language texts |
| US20120010872A1 (en) * | 2006-10-10 | 2012-01-12 | Abbyy Software Ltd | Method and System for Semantic Searching |
| CN101782897A (en) * | 2010-03-17 | 2010-07-21 | 上海大学 | Chinese corpus labeling method based on events |
| CN101937430A (en) * | 2010-09-03 | 2011-01-05 | 清华大学 | Method for extracting event sentence pattern from Chinese sentence |
| CN101957812A (en) * | 2010-09-21 | 2011-01-26 | 上海大学 | Verb semantic information extracting method based on event ontology |
Non-Patent Citations (3)
| Title |
|---|
| 付剑锋等: "《基于层叠条件随机场的事件因果关系抽取》", 《模式识别与人工智能》, vol. 24, no. 4, 30 August 2011 (2011-08-30), pages 567 - 573 * |
| 刘宗田等: "《面向事件的本体研究》", 《计算机科学》, vol. 36, no. 11, 30 November 2009 (2009-11-30), pages 189 - 199 * |
| 朱怀: "《事件结构理论的起源与发展》", 《外语学刊》, no. 163, 31 December 2011 (2011-12-31), pages 82 - 85 * |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104090867B (en) * | 2014-07-17 | 2016-09-21 | 北京中电拓方科技股份有限公司 | A kind of method performing event based on Mining Security Quality standard |
| CN104090867A (en) * | 2014-07-17 | 2014-10-08 | 北京中电拓方科技发展有限公司 | Method for executing event based on coal mine safety quality standard |
| CN105243056A (en) * | 2015-09-07 | 2016-01-13 | 饶志刚 | Punctuation mark processing based Chinese syntax analysis method and apparatus |
| CN105243056B (en) * | 2015-09-07 | 2018-02-06 | 饶志刚 | A kind of Chinese parsing method and device based on punctuation mark processing |
| CN105224640A (en) * | 2015-09-25 | 2016-01-06 | 杭州朗和科技有限公司 | A kind of method and apparatus extracting viewpoint |
| CN106021234A (en) * | 2016-05-31 | 2016-10-12 | 徐子涵 | Label extraction method and system |
| CN107451158B (en) * | 2016-06-01 | 2021-01-19 | 中国科学院地理科学与资源研究所 | Method for extracting semantic roles of traffic events in web text |
| CN107451158A (en) * | 2016-06-01 | 2017-12-08 | 中国科学院地理科学与资源研究所 | Traffic events semantic role abstracting method in a kind of network text |
| CN106897364B (en) * | 2017-01-12 | 2021-02-23 | 上海大学 | Chinese reference corpus construction method based on events |
| CN106897364A (en) * | 2017-01-12 | 2017-06-27 | 上海大学 | Chinese based on event refers to building of corpus method |
| CN107679035B (en) * | 2017-10-11 | 2020-06-12 | 石河子大学 | Information intention detection method, device, equipment and storage medium |
| CN107679035A (en) * | 2017-10-11 | 2018-02-09 | 石河子大学 | A kind of information intent detection method, device, equipment and storage medium |
| CN108197294B (en) * | 2018-01-22 | 2021-10-22 | 桂林电子科技大学 | Text automatic generation method based on deep learning |
| CN108197294A (en) * | 2018-01-22 | 2018-06-22 | 桂林电子科技大学 | A kind of text automatic generation method based on deep learning |
| CN108595421A (en) * | 2018-04-13 | 2018-09-28 | 北京神州泰岳软件股份有限公司 | A kind of abstracting method, the apparatus and system of Chinese entity associated relationship |
| CN108595421B (en) * | 2018-04-13 | 2022-04-08 | 鼎富智能科技有限公司 | Method, device and system for extracting Chinese entity association relationship |
| CN108959464A (en) * | 2018-06-19 | 2018-12-07 | 李勤骞 | Learning method and system containing auxiliary word |
| CN108959464B (en) * | 2018-06-19 | 2021-06-08 | 李勤骞 | Learning method and system containing auxiliary words |
| CN110781369A (en) * | 2018-07-11 | 2020-02-11 | 天津大学 | Emotional cause mining method based on dependency syntax and generalized causal network |
| CN110851560A (en) * | 2018-07-27 | 2020-02-28 | 杭州海康威视数字技术股份有限公司 | Information retrieval method, device and equipment |
| CN109471936A (en) * | 2018-10-11 | 2019-03-15 | 上海叔本华智能科技有限公司 | A kind of method and system for plant maintenance information progress tagsort |
| CN109815481A (en) * | 2018-12-17 | 2019-05-28 | 北京百度网讯科技有限公司 | Method, apparatus, equipment and the computer storage medium of event extraction are carried out to text |
| CN110874531A (en) * | 2020-01-20 | 2020-03-10 | 湖南蚁坊软件股份有限公司 | Topic analysis method and device and storage medium |
| CN111581954A (en) * | 2020-05-15 | 2020-08-25 | 中国人民解放军国防科技大学 | Text event extraction method and device based on grammar dependency information |
| CN112148838B (en) * | 2020-09-23 | 2024-04-19 | 北京中电普华信息技术有限公司 | Service source object extraction method and device |
| CN112148838A (en) * | 2020-09-23 | 2020-12-29 | 北京中电普华信息技术有限公司 | Business source object extraction method and device |
| CN112699664A (en) * | 2021-01-08 | 2021-04-23 | 中国专利信息中心 | Chinese syntax analysis method and system |
| CN115017913A (en) * | 2022-04-21 | 2022-09-06 | 广州世纪华轲科技有限公司 | Semantic component analysis method based on master-slave framework mode |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103268311A (en) | Event-structure-based Chinese statement analysis method | |
| Bamman et al. | An annotated dataset of coreference in English literature | |
| Cunningham | Information extraction, automatic | |
| Morris et al. | Lexical cohesion computed by thesaural relations as an indicator of the structure of text | |
| Fillmore | Border conflicts: FrameNet meets construction grammar | |
| Ananiadou et al. | The English language in the digital age | |
| Lyngfelt et al. | Adding a constructicon to the Swedish resource network of Språkbanken. | |
| Øvrelid et al. | Syntactic scope resolution in uncertainty analysis | |
| CN111581953A (en) | Method for automatically analyzing grammar phenomenon of English text | |
| Zikánová et al. | Discourse and coherence | |
| Spyns et al. | Essential speech and language technology for Dutch | |
| Khorjuvenkar et al. | Parts of speech tagging for Konkani language | |
| CN110457691A (en) | Feeling curve analysis method and device based on drama role | |
| Zhang et al. | Semi-automatic emotion recognition from textual input based on the constructed emotion thesaurus | |
| Arivazhagan et al. | Labeling the semantic roles of commas | |
| El-Najjar et al. | Improving dependency parsing of verbal arabic sentences using semantic features | |
| Ghosh et al. | Clause identification and classification in bengali | |
| Cullip | Text technology: The power-tool of grammatical metaphor | |
| Borrega et al. | What do we mean when we speak about Named Entities | |
| Agnoloni et al. | Semantic processing of legal texts | |
| Mititelu et al. | Improving parsing using morpho-syntactic and semantic information | |
| Intasaw et al. | Basic principles for segmenting Thai EDUs | |
| Sharma | Sentence Reduction for Syntactic Analysis of Compound Sentences in Punjabi Language | |
| Vileiniškis et al. | Searching the web by meaning: a case study of Lithuanian news websites | |
| Forsbom | Rhetorical structure theory in natural language generation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130828 |