CN111858840A - Intention identification method and device based on concept graph - Google Patents
Intention identification method and device based on concept graph Download PDFInfo
- Publication number
- CN111858840A CN111858840A CN201910329359.6A CN201910329359A CN111858840A CN 111858840 A CN111858840 A CN 111858840A CN 201910329359 A CN201910329359 A CN 201910329359A CN 111858840 A CN111858840 A CN 111858840A
- Authority
- CN
- China
- Prior art keywords
- intention
- question
- concept
- user
- keyword
- 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
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000013507 mapping Methods 0.000 claims description 14
- 238000000605 extraction Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 abstract description 29
- 230000003993 interaction Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 5
- 238000006243 chemical reaction Methods 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000005034 decoration Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/3332—Query translation
- G06F16/3334—Selection or weighting of terms from queries, including natural language queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Artificial Intelligence (AREA)
- User Interface Of Digital Computer (AREA)
Abstract
The invention belongs to the field of intention identification, and discloses an intention identification method and device based on a concept graph, wherein the method comprises the following steps: collecting voice information of a user; carrying out grammar analysis on the voice information, and extracting question words and key words in the voice information; matching the question words with nodes in a question concept graph to obtain matched nodes, wherein the question concept graph is constructed in advance; and acquiring the intention of the user according to the intention concept corresponding to the matched node and the keyword. According to the invention, the method generates the question concept diagrams from various arbitrary questions, so that the user intention can be more easily and accurately acquired when the user intention is acquired in the generated question concept diagrams according to the question words in the voice information, and the search and identification speed is higher due to clear hierarchy of various questions in the question concept diagrams.
Description
Technical Field
The invention belongs to the technical field of semantic recognition, and particularly relates to an intention recognition method and device based on a concept graph.
Background
With the rapid development of intelligent terminals and network technologies, people are more and more accustomed to using intelligent terminals to fulfill various requirements. Natural language has gradually become the most mainstream man-machine interaction mode in the field of intelligent services as the most convenient and natural way for human to express self thought.
In a human-computer interaction scene, intention recognition is an essential link, the intention of a user is known mainly by analyzing voice input by the user, the voice is converted into a structured data format which can be understood by a machine, and then corresponding feedback is made, so that in the human-computer interaction scene, accurate recognition of the intention of the user is the basis for making correct response. For students in lower grades, incomplete language expression, randomness and the like may occur in the language expression process due to the language learning stage. Therefore, in the voice electronic products used by students, if a conventional intention recognition method is used, the intention recognition is prone to be inaccurate, so that the intelligent terminal cannot make correct responses, and the use experience of the user is affected.
Disclosure of Invention
The invention aims to provide an intention identification method and device based on a concept graph, which are used for generating a question concept graph by various optional questions so as to more easily and accurately acquire user intention.
The technical scheme provided by the invention is as follows:
in one aspect, a concept graph-based intention identification method is provided, including:
collecting voice information of a user;
Carrying out grammar analysis on the voice information, and extracting question words and key words in the voice information;
matching the question words with nodes in a question concept graph to obtain matched nodes, wherein the question concept graph is constructed in advance;
and acquiring the intention of the user according to the intention concept corresponding to the matched node and the keyword.
Further preferably, the construction method of the concept map of the method of inquiry comprises the following steps:
collecting a large amount of user corpora;
extracting the question words in the corpus;
acquiring the intention concepts of the question words with the same intention in the corpus;
establishing a mapping relation between the question words and the intention concepts;
and constructing a question concept graph according to the mapping relation between the question words and the intention concepts.
Further preferably, the obtaining the intention of the user according to the intention concept corresponding to the matched node and the keyword specifically includes:
searching a target word matched with the keyword from a preset word bank;
determining the semantics of the keywords according to the semantics of the target words;
and acquiring the intention of the user according to the intention concept corresponding to the matched node and the semantics of the keyword.
Further preferably, after parsing the voice information and extracting the question words and the keywords in the voice information, the method further includes:
when the keywords contain specific keywords, receiving a touch signal of the user;
acquiring image information of the user selection area according to the touch signal;
the obtaining the intention of the user according to the intention concept corresponding to the matched node and the keyword specifically includes:
and acquiring the intention of the user according to the intention concept corresponding to the matched node, the keyword and the image information.
Further preferably, the obtaining the intention of the user according to the intention concept corresponding to the matched node, the keyword, and the image information specifically includes:
identifying character information in the image information;
replacing the text information with a specific keyword in the keywords;
and acquiring the intention of the user according to the intention concept corresponding to the matched node, the keyword and the text information.
In another aspect, an intention recognition apparatus based on a concept graph is also provided, including:
the acquisition module is used for acquiring voice information of a user;
The extraction module is used for carrying out grammar analysis on the voice information and extracting the question words and the key words in the voice information;
the matching module is used for matching the question words with nodes in a question concept graph to obtain matched nodes, wherein the question concept graph is constructed in advance;
and the intention acquisition module is used for acquiring the intention of the user according to the intention concept corresponding to the matched node and the keyword.
Further preferably, the system also comprises a construction module;
the building module comprises:
the corpus collection unit is used for collecting a large amount of user corpuses;
the word extraction unit is used for extracting the question words in the corpus;
the concept acquisition unit is used for acquiring the intention concepts of the same-intention question words in the corpus;
the relationship establishing unit is used for establishing a mapping relationship between the question and legal terms and the intention concepts;
and the construction unit is used for constructing a question and law concept graph according to the mapping relation between the question and law words and the intention concepts.
Further preferably, the intention acquisition module includes:
the word searching unit is used for searching a target word matched with the keyword from a preset word bank;
The semantic determining unit is used for determining the semantics of the keywords according to the semantics of the target words;
and the intention acquisition unit is used for acquiring the intention of the user according to the intention concept corresponding to the matched node and the semantics of the keyword.
Further preferably, the method further comprises the following steps:
the signal receiving module is used for receiving a touch signal of the user when the keyword contains a specific keyword;
the image acquisition module is used for acquiring the image information of the user selection area according to the touch signal;
the intention acquisition module is further used for acquiring the intention of the user according to the intention concept corresponding to the matched node, the keyword and the image information.
Further preferably, the intention acquisition module includes:
the identification unit is used for identifying character information in the image information;
the replacing unit is used for replacing the text information with a specific keyword in the keywords;
and the intention acquisition unit is used for acquiring the intention of the user according to the intention concept corresponding to the matched node, the keyword and the character information.
Compared with the prior art, the intention identification method and device based on the concept graph have the advantages that: according to the invention, the method generates the question concept diagrams from various arbitrary questions, so that the user intention can be more easily and accurately acquired when the user intention is acquired in the generated question concept diagrams according to the question words in the voice information, and the search and identification speed is higher due to clear hierarchy of various questions in the question concept diagrams.
Drawings
The above features, technical features, advantages and implementations of a method and apparatus for intent recognition based on conceptual views will be further described in the following detailed description of preferred embodiments in a clearly understandable manner with reference to the accompanying drawings.
FIG. 1 is a schematic flow chart diagram of a concept graph-based intention recognition method according to a first embodiment of the present invention;
FIG. 2 is a schematic flow chart diagram of a concept graph-based intention recognition method according to a second embodiment of the present invention;
FIG. 3 is a schematic flow chart diagram of a concept graph-based intention recognition method according to a third embodiment of the present invention;
FIG. 4 is a schematic flow chart diagram of a fourth embodiment of the concept graph-based intention recognition method of the present invention;
FIG. 5 is a schematic flow chart diagram of a fifth embodiment of the intent recognition method based on concept graph according to the present invention;
FIG. 6 is a block diagram schematically illustrating the structure of an embodiment of a concept-based intention recognition apparatus according to the present invention;
FIG. 7 is a block diagram schematically illustrating the structure of another embodiment of the concept-diagram-based intention recognition apparatus according to the present invention.
Description of the reference numerals
100. An acquisition module; 200. an extraction module; 300. a matching module; 400. an intention acquisition module; 410. a word search unit; 420. a semantic determination unit; 430. an intention acquisition unit; 440. an identification unit; 450. a replacement unit; 500. building a module; 510. a corpus collection unit; 520. a word extraction unit; 530. a concept acquisition unit; 540. a relationship establishing unit; 550. a building unit; 600. a signal receiving module; 700. and an image acquisition module.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following description will be made with reference to the accompanying drawings. It is obvious that the drawings in the following description are only some examples of the invention, and that for a person skilled in the art, other drawings and embodiments can be derived from them without inventive effort.
For the sake of simplicity, the drawings only schematically show the parts relevant to the present invention, and they do not represent the actual structure as a product. In addition, in order to make the drawings concise and understandable, components having the same structure or function in some of the drawings are only schematically illustrated or only labeled. In this document, "one" means not only "only one" but also a case of "more than one".
In the invention, the intention recognition is carried out through the concept graph, so that the accuracy rate of the intention recognition can be improved; the intention identification method based on the concept graph can be applied to intelligent terminal equipment; for example: for convenience of understanding, the family education machine is used as a subject for explanation in the following embodiments, but those skilled in the art should understand that the intention identification method based on the concept diagram can also be applied to other intelligent terminal devices as long as the corresponding functions can be realized.
According to a first embodiment provided by the present invention, as shown in fig. 1, an intent recognition method based on a concept graph includes:
s100, collecting voice information of a user;
in particular, the voice of the user may be collected by a microphone or other voice collection device. The microphone and other voice collecting devices can be built-in of the family education machine and can also be external equipment. The voice information may be voice input by the user in real time.
S200, performing grammar analysis on the voice information, and extracting question words and key words in the voice information;
specifically, after the voice information is acquired, the voice information may be converted into text information, and then the converted text information is subjected to word segmentation and syntax analysis to extract the question words and the keywords from the text information. The conversion of the voice information into text can be realized by various existing technical means, and the invention is not limited to the use of a certain conversion method, and is not described in detail herein.
The question words here refer to words representing various questions in the voice interaction process, such as "how to do", "how to understand", "I do not understand", and the like. The keyword here refers to a noun having a practical meaning in the voice message, i.e. a noun having a corresponding entity; for example, "language video" in "where the language video is" is the keyword, "join and remove problem" in "how do join and remove problem" is the keyword, "and" Colosseus "in Colosseus" is the keyword.
S300, matching the question words with nodes in a question concept graph to obtain matched nodes, wherein the question concept graph is constructed in advance;
specifically, the concept graph is a graph method for representing concepts by nodes and representing relationships among the concepts by connecting lines. A conceptual graph generally consists of "nodes," links, "and" related text labels. The query concept graph is a concept graph generated by using different queries according to the relationship between the queries. In the conceptual diagram of the method, a root node includes a plurality of child nodes, and a child node is a lower concept of the root node.
For example: the root node of the concept graph of the method comprises a plurality of concepts of the method, such as the method for expressing 'explanation', the method for expressing 'searching', the method for expressing 'adjustment' and the like.
The different concepts of the question method include a plurality of question methods, for example, the question method for indicating "explanation" includes "don't understand", "how to do", "what", and the like. The question of "find" includes "where", etc.
And matching the question words extracted from the language information with the nodes in the question concept graph, namely searching the nodes matched with the question words in the question concept graph.
S400, according to the intention concept corresponding to the matched node and the keyword, the intention of the user is obtained.
Specifically, after the matched nodes are obtained, the corresponding intention concepts can be obtained according to the matched nodes, and then the intention of the user is obtained according to the corresponding intention concepts and by combining the keywords.
Illustratively, the voice information is "what the unary linear equation is", the question word extracted from the voice information is "what" and the extracted keyword is "the unary linear equation", the node matched by the question word in the question concept graph is "what", and the intention concept of the node in the question concept graph is "explanation", and the keyword "the unary linear equation" and the "explanation" are combined to obtain the content related to the user's intention for searching the unary linear equation, such as the explanation for searching the unary linear equation, the exercise of the unary linear equation, and the like.
In the embodiment, the method and the device generate the query concept diagrams from various arbitrary queries, so that when the user intention is acquired in the generated query concept diagrams according to the query words in the voice information, the user intention is more easily and accurately acquired, and in the query concept diagrams, various query hierarchies are clear, so that the searching and identifying speed is higher.
According to a second embodiment provided by the present invention, as shown in fig. 2, an intention identification method based on a concept graph is provided, and in this embodiment, on the basis of the first embodiment, a construction method of the question-and-law concept graph includes:
s010 collects a large amount of user corpora;
specifically, in the process that the user uses the family education machine, various linguistic data generated when the user performs voice interaction with the family education machine are collected.
S020 extracting question words in the corpus;
specifically, after a large amount of corpora are collected, the corpora may be converted into text information, then the converted text information is subjected to word segmentation and grammar analysis, and then the question words related to the question are extracted from the converted text information.
For example, the corpora are "how do to connect with the problem", "do not move with the graphic", "do not know with the semantic video", "help us open the Chinese practice problem", "volume too small", "volume too large", "font too small", "do you want to pre-study the Chinese", "open the historical textbook", etc. The question words extracted from the corpus are "how to do", "I don't understand", "where", "open", "too small", "too large", "preview", etc.
S030 obtains the intention concepts of the equivocals with the same intention in the corpus;
specifically, after various question and grammar words are extracted from the corpus, the question and grammar words are classified, the question and grammar words with the same intention are classified into one class, and then the intention concepts of the question and grammar words with the same intention are obtained.
For example, the words "how to do", "do not do" and "do not understand" all indicate that a certain content is not understood, cannot be answered, is unknown, and the like, and a home education machine is required to search and display corresponding answers, subject explanation, or lectures, and the like; therefore, the words of inquiry such as "how to do", "do not understand" and the like can be classified into one category.
The words "where", "open" and "preview" all represent that a certain content needs to be found and the corresponding content needs to be opened by the family education machine, so that the words "where", "open" and "preview" can be classified into one category.
The words "too small" and "too large" indicate that a certain object needs to be adjusted, such as adjusting the volume, adjusting the font size, adjusting the video display size, and the like; thus, the words "too small" and "too large" can be grouped together.
After classifying all the obtained question and grammar words, determining the intention concept of each classified question and grammar word according to the intention expressed by the question and grammar words; if the intention concepts of the question words "how to do", "I don't", "I do not understand" can be adjusted to "explain", the intention concepts of the question words "where", "open", "preview" are adjusted to "find object", and the intention concepts of the question words too small and too large are adjusted to "adjust".
S040 establishes a mapping relation between the question and legal terms and the intention concepts;
specifically, after the question words are classified and the intention concept of each class of question words is determined, a mapping relation between the question words and the intention concept is established, that is, a corresponding relation between the intention concept and the question words is established.
S050 according to the mapping relation between the question words and the intention concepts, a question concept graph is constructed.
Specifically, according to the established correspondence between the question words and the intention concepts, a question concept graph can be constructed.
In this embodiment, a large number of arbitrary spoken interrogations can be collected by collecting the corpus of the user when using the intelligent terminal device, and then the arbitrary interrogations are classified and sorted to generate the interrogations concept diagram, so that the recognition rate is high when intention recognition is performed through the interrogations concept diagram.
According to a third embodiment provided by the present invention, as shown in fig. 3, an intention identifying method based on a concept graph includes:
s100, collecting voice information of a user;
s200, performing grammar analysis on the voice information, and extracting question words and key words in the voice information;
s300, matching the question words with nodes in a question concept graph to obtain matched nodes, wherein the question concept graph is constructed in advance;
S410, searching a target word matched with the keyword from a preset word stock;
specifically, when the keyword intention is obtained, a word bank can be established first, the keywords in the word bank are nominal words, the keywords in the word bank can be crawled from the network through a crawler technology, and the semantics of the words in the word bank can be obtained. And then matching the keywords extracted from the voice information with a word bank in the word bank to find out a matched target word.
S420, determining the semantics of the keywords according to the semantics of the target words;
specifically, after the target word is found, the semantics of the keyword in the voice information are determined according to the semantics corresponding to the target word, that is, the semantics of the keyword in the voice information are determined, and the semantic is converted into a structured data format which can be understood by a machine.
S430, according to the intention concept corresponding to the matched node and the semantics of the keyword, the intention of the user is obtained.
Specifically, after the semantics of the keyword and the intention concept are obtained, the intention of the user can be obtained according to the intention concept and the semantics of the keyword.
For example, the voice information is "how to do the continuous removal problem", the keyword extracted from the voice information is "how to do the continuous removal problem", the question concept corresponding to the node matched with the question word "how to do" is "explained" according to the question concept diagram, the semantic meaning of the keyword "continuous removal problem" is "continuous removal problem", and the intention of the user is "continuous removal problem" according to the question concept "explanation" and the keyword "continuous removal problem".
In the embodiment, the word bank is established, and the semantics of the keywords are obtained by the keyword matching method, so that the method is more convenient and quicker compared with other semantic identification methods.
According to a fourth embodiment provided by the present invention, as shown in fig. 4, an intention identifying method based on a concept graph includes:
s100, collecting voice information of a user;
s200, performing grammar analysis on the voice information, and extracting question words and key words in the voice information;
s210, when the keywords contain specific keywords, receiving a touch signal of the user;
specifically, the specific keyword means "the word", "the subject", "the word", "the", and the like refer to pronouns. And when the voice information contains the specific keyword, receiving a touch signal of the user on the touch screen. The touch signal may be a continuous sliding touch signal of the user on the touch screen, such as a straight line formed by continuous touch, or a circular frame, an oval frame, an irregular graphic frame, etc. formed by continuous touch.
S220, acquiring image information of the user selection area according to the touch signal;
specifically, after receiving the touch signal, the image information of the user selection area may be obtained according to the touch signal, for example, after a straight line is formed by continuous touch, a rectangular frame is constructed by using the straight line as a diagonal line, and an image in the rectangular frame is captured, where the image in the rectangular frame is the image information of the user selection area.
And if a circular frame, an oval frame or the like is formed by continuous touch, intercepting an image in the formed frame, wherein the image in the frame is the image information of the area selected by the user.
Here, the touch signal is limited to a continuous touch signal, so that an error operation can be prevented, the occurrence of an error rate can be reduced, for example, a user can be prevented from forming two point touches on the touch screen due to the error operation, and an image in a frame formed by the two points can be acquired.
When the voice information contains the specific keyword, it is described that the voice information input by the user is missing, for example, "how to solve the question", the family education machine does not know what the question is, and therefore, even after performing semantic analysis on the voice information, the accurate intention of the user cannot be obtained, and therefore, when it is determined that the voice information contains the specific keyword, it is necessary to further obtain image information pointed by a finger of the user or in a frame selection area to obtain information of the question.
S300, matching the question words with nodes in a question concept graph to obtain matched nodes, wherein the question concept graph is constructed in advance;
s440, according to the intention concept corresponding to the matched node, the keyword and the image information, the intention of the user is obtained.
Specifically, according to the obtained intention concept and the semantics of the keywords, the obtained image information is combined, and the three are fused, so that the real intention of the user can be obtained, and corresponding feedback is given.
According to a fifth embodiment provided by the present invention, as shown in fig. 5, an intention identifying method based on a concept graph includes:
s100, collecting voice information of a user;
s200, performing grammar analysis on the voice information, and extracting question words and key words in the voice information;
s210, when the keywords contain specific keywords, receiving a touch signal of the user;
s220, acquiring image information of the user selection area according to the touch signal;
s300, matching the question words with nodes in a question concept graph to obtain matched nodes, wherein the question concept graph is constructed in advance;
s441 identifies text information in the image information;
specifically, after image information is obtained, the image is processed and identified to obtain character information in the image information, where the character information is information of a certain topic or information of a certain character.
S442 replacing the text message with a specific keyword among the keywords;
S443, obtaining the intention of the user according to the intention concept corresponding to the matched node, the keyword and the character information.
Specifically, the acquired text information is replaced with the specific keyword, that is, the topic is replaced with the acquired text information. And then acquiring the real intention of the user according to the intention concept corresponding to the matched node, the semantics of the rest keywords and the acquired character information.
For example, when a user views an extracurricular reading on a home teaching machine, when "Colosseus" is seen, the "Colosseus" does not know and seek help, and then the user says "what these words are read", the home teaching machine acquires the voice information and judges that the voice information contains a specific keyword, the image information of the user-selected area is acquired based on the touch signal, for example, the image of "Colosseus" is acquired, the character information of "Colosseus" is acquired by performing character recognition on the image, and then the "what is read" is combined with the question word, and the true semantic of the user is "what is read" in Colosseus ". Therefore, words unknown to the user or questions which cannot be solved can be helped through the family education machine so as to assist the user in learning.
According to a sixth embodiment provided by the present invention, as shown in fig. 6, an intention identifying apparatus based on a concept graph includes:
the acquisition module 100 is used for acquiring voice information of a user;
in particular, the voice of the user may be collected by a microphone or other voice collection device. The microphone and other voice collecting devices can be built-in of the family education machine and can also be external equipment. The voice information may be voice input by the user in real time.
An extracting module 200, configured to perform syntax analysis on the voice information, and extract a question word and a keyword in the voice information;
specifically, after the voice information is acquired, the voice information may be converted into text information, and then the converted text information is subjected to word segmentation and syntax analysis to extract the question words and the keywords from the text information. The conversion of the voice information into text can be realized by various existing technical means, and the invention is not limited to the use of a certain conversion method, and is not described in detail herein.
The question words here refer to words representing various questions in the voice interaction process, such as "how to do", "how to understand", "I do not understand", and the like. The keyword here refers to a noun having a practical meaning in the voice message, i.e. a noun having a corresponding entity; for example, "language video" in "where the language video is" is the keyword, "join and remove problem" in "how do join and remove problem" is the keyword, "and" Colosseus "in Colosseus" is the keyword.
The matching module 300 is configured to match the question words with nodes in a question concept graph to obtain matched nodes, where the question concept graph is pre-constructed;
specifically, the concept graph is a graph method for representing concepts by nodes and representing relationships among the concepts by connecting lines. A conceptual graph generally consists of "nodes," links, "and" related text labels. The query concept graph is a concept graph generated by using different queries according to the relationship between the queries. In the conceptual diagram of the method, a root node includes a plurality of child nodes, and a child node is a lower concept of the root node.
For example: the root node of the concept graph of the method comprises a plurality of concepts of the method, such as the method for expressing 'explanation', the method for expressing 'searching', the method for expressing 'adjustment' and the like.
The different concepts of the question method include a plurality of question methods, for example, the question method for indicating "explanation" includes "don't understand", "how to do", "what", and the like. The question of "find" includes "where", etc.
And matching the question words extracted from the language information with the nodes in the question concept graph, namely searching the nodes matched with the question words in the question concept graph.
An intention obtaining module 400, configured to obtain the intention of the user according to the intention concept corresponding to the matched node and the keyword.
Specifically, after the matched nodes are obtained, the corresponding intention concepts can be obtained according to the matched nodes, and then the intention of the user is obtained according to the corresponding intention concepts and by combining the keywords.
Illustratively, the voice information is "what the unary linear equation is", the question word extracted from the voice information is "what" and the extracted keyword is "the unary linear equation", the node matched by the question word in the question concept graph is "what", and the intention concept of the node in the question concept graph is "explanation", and the keyword "the unary linear equation" and the "explanation" are combined to obtain the content related to the user's intention for searching the unary linear equation, such as the explanation for searching the unary linear equation, the exercise of the unary linear equation, and the like.
In the embodiment, the method and the device generate the query concept diagrams from various arbitrary queries, so that when the user intention is acquired in the generated query concept diagrams according to the query words in the voice information, the user intention is more easily and accurately acquired, and in the query concept diagrams, various query hierarchies are clear, so that the searching and identifying speed is higher.
Preferably, a building module 500 is also included;
the building block 500 comprises:
a corpus collection unit 510 for collecting a large amount of user corpuses;
specifically, in the process that the user uses the family education machine, various linguistic data generated when the user performs voice interaction with the family education machine are collected.
A word extraction unit 520, configured to extract the question words in the corpus;
specifically, after a large amount of corpora are collected, the corpora may be converted into text information, then the converted text information is subjected to word segmentation and grammar analysis, and then the question words related to the question are extracted from the converted text information.
For example, the corpora are "how do to connect with the problem", "do not move with the graphic", "do not know with the semantic video", "help us open the Chinese practice problem", "volume too small", "volume too large", "font too small", "do you want to pre-study the Chinese", "open the historical textbook", etc. The question words extracted from the corpus are "how to do", "I don't understand", "where", "open", "too small", "too large", "preview", etc.
A concept obtaining unit 530, configured to obtain the intention concepts of the equivocational interrogative words with the same intention in the corpus;
Specifically, after various question and grammar words are extracted from the corpus, the question and grammar words are classified, the question and grammar words with the same intention are classified into one class, and then the intention concepts of the question and grammar words with the same intention are obtained.
For example, the words "how to do", "do not do" and "do not understand" all indicate that a certain content is not understood, cannot be answered, is unknown, and the like, and a home education machine is required to search and display corresponding answers, subject explanation, or lectures, and the like; therefore, the words of inquiry such as "how to do", "do not understand" and the like can be classified into one category.
The words "where", "open" and "preview" all represent that a certain content needs to be found and the corresponding content needs to be opened by the family education machine, so that the words "where", "open" and "preview" can be classified into one category.
The words "too small" and "too large" indicate that a certain object needs to be adjusted, such as adjusting the volume, adjusting the font size, adjusting the video display size, and the like; thus, the words "too small" and "too large" can be grouped together.
After classifying all the obtained question and grammar words, determining the intention concept of each classified question and grammar word according to the intention expressed by the question and grammar words; if the intention concepts of the question words "how to do", "I don't", "I do not understand" can be adjusted to "explain", the intention concepts of the question words "where", "open", "preview" are adjusted to "find object", and the intention concepts of the question words too small and too large are adjusted to "adjust".
A relationship establishing unit 540, configured to establish a mapping relationship between the question and legal terms and the intent concept;
specifically, after the question words are classified and the intention concept of each class of question words is determined, a mapping relation between the question words and the intention concept is established, that is, a corresponding relation between the intention concept and the question words is established.
The constructing unit 550 is configured to construct a concept graph according to the mapping relationship between the question words and the intention concepts.
Specifically, according to the established correspondence between the question words and the intention concepts, a question concept graph can be constructed.
In this embodiment, a large number of arbitrary spoken interrogations can be collected by collecting the corpus of the user when using the intelligent terminal device, and then the arbitrary interrogations are classified and sorted to generate the interrogations concept diagram, so that the recognition rate is high when intention recognition is performed through the interrogations concept diagram.
Preferably, the intention acquisition module 400 includes:
a word searching unit 410, configured to search a preset word bank for a target word matching the keyword;
specifically, when the keyword intention is obtained, a word bank can be established first, the keywords in the word bank are nominal words, the keywords in the word bank can be crawled from the network through a crawler technology, and the semantics of the words in the word bank can be obtained. And then matching the keywords extracted from the voice information with a word bank in the word bank to find out a matched target word.
A semantic determining unit 420, configured to determine semantics of the keyword according to the semantics of the target word;
specifically, after the target word is found, the semantics of the keyword in the voice information are determined according to the semantics corresponding to the target word, that is, the semantics of the keyword in the voice information are determined, and the semantic is converted into a structured data format which can be understood by a machine.
An intention obtaining unit 430, configured to obtain the intention of the user according to the intention concept corresponding to the matched node and the semantics of the keyword.
Specifically, after the semantics of the keyword and the intention concept are obtained, the intention of the user can be obtained according to the intention concept and the semantics of the keyword.
For example, the voice information is "how to do the continuous removal problem", the keyword extracted from the voice information is "how to do the continuous removal problem", the question concept corresponding to the node matched with the question word "how to do" is "explained" according to the question concept diagram, the semantic meaning of the keyword "continuous removal problem" is "continuous removal problem", and the intention of the user is "continuous removal problem" according to the question concept "explanation" and the keyword "continuous removal problem".
In the embodiment, the word bank is established, and the semantics of the keywords are obtained by the keyword matching method, so that the method is more convenient and quicker compared with other semantic identification methods.
According to a seventh embodiment provided by the present invention, as shown in fig. 7, an intention identifying apparatus based on a concept graph includes:
the acquisition module 100 is used for acquiring voice information of a user;
an extracting module 200, configured to perform syntax analysis on the voice information, and extract a question word and a keyword in the voice information;
the matching module 300 is configured to match the question words with nodes in a question concept graph to obtain matched nodes, where the question concept graph is pre-constructed;
an intention obtaining module 400, configured to obtain the intention of the user according to the intention concept corresponding to the matched node and the keyword.
Preferably, the method further comprises the following steps:
a signal receiving module 600, configured to receive a touch signal of the user when the keyword includes a specific keyword;
specifically, the specific keyword means "the word", "the subject", "the word", "the", and the like refer to pronouns. And when the voice information contains the specific keyword, receiving a touch signal of the user on the touch screen. The touch signal may be a continuous sliding touch signal of the user on the touch screen, such as a straight line formed by continuous touch, or a circular frame, an oval frame, an irregular graphic frame, etc. formed by continuous touch.
An image obtaining module 700, configured to obtain image information of the user selected area according to the touch signal;
specifically, after receiving the touch signal, the image information of the user selection area may be obtained according to the touch signal, for example, after a straight line is formed by continuous touch, a rectangular frame is constructed by using the straight line as a diagonal line, and an image in the rectangular frame is captured, where the image in the rectangular frame is the image information of the user selection area.
And if a circular frame, an oval frame or the like is formed by continuous touch, intercepting an image in the formed frame, wherein the image in the frame is the image information of the area selected by the user.
Here, the touch signal is limited to a continuous touch signal, so that an error operation can be prevented, the occurrence of an error rate can be reduced, for example, a user can be prevented from forming two point touches on the touch screen due to the error operation, and an image in a frame formed by the two points can be acquired.
When the voice information contains the specific keyword, it is described that the voice information input by the user is missing, for example, "how to solve the question", the family education machine does not know what the question is, and therefore, even after performing semantic analysis on the voice information, the accurate intention of the user cannot be obtained, and therefore, when it is determined that the voice information contains the specific keyword, it is necessary to further obtain image information pointed by a finger of the user or in a frame selection area to obtain information of the question.
The intention obtaining module 400 is further configured to obtain the intention of the user according to the intention concept corresponding to the matched node, the keyword, and the image information.
Specifically, according to the obtained intention concept and the semantics of the keywords, the obtained image information is combined, and the three are fused, so that the real intention of the user can be obtained, and corresponding feedback is given.
Preferably, the intention acquisition module 400 includes:
an identifying unit 440, configured to identify text information in the image information;
a replacing unit 450, configured to replace the text information with a specific keyword in the keywords;
an intention obtaining unit 430, configured to obtain the intention of the user according to the intention concept corresponding to the matched node, the keyword, and the text information.
Specifically, after image information is obtained, the image is processed and identified to obtain character information in the image information, where the character information is information of a certain topic or information of a certain character.
And replacing the acquired character information with the specific keyword, namely replacing the question with the acquired character information. And then acquiring the real intention of the user according to the intention concept corresponding to the matched node, the semantics of the rest keywords and the acquired character information.
For example, when a user views an extracurricular reading on a home teaching machine, when "Colosseus" is seen, the "Colosseus" does not know and seek help, and then the user says "what these words are read", the home teaching machine acquires the voice information and judges that the voice information contains a specific keyword, the image information of the user-selected area is acquired based on the touch signal, for example, the image of "Colosseus" is acquired, the character information of "Colosseus" is acquired by performing character recognition on the image, and then the "what is read" is combined with the question word, and the true semantic of the user is "what is read" in Colosseus ". Therefore, words unknown to the user or questions which cannot be solved can be helped through the family education machine so as to assist the user in learning.
It should be noted that the above embodiments can be freely combined as necessary. The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (10)
1. An intention recognition method based on a concept graph is characterized by comprising the following steps:
Collecting voice information of a user;
carrying out grammar analysis on the voice information, and extracting question words and key words in the voice information;
matching the question words with nodes in a question concept graph to obtain matched nodes, wherein the question concept graph is constructed in advance;
and acquiring the intention of the user according to the intention concept corresponding to the matched node and the keyword.
2. The method for identifying an intention based on a concept graph according to claim 1, wherein the method for constructing the concept graph of the question method comprises the following steps:
collecting a large amount of user corpora;
extracting the question words in the corpus;
acquiring the intention concepts of the question words with the same intention in the corpus;
establishing a mapping relation between the question words and the intention concepts;
and constructing a question concept graph according to the mapping relation between the question words and the intention concepts.
3. The method as claimed in claim 1, wherein the obtaining the user's intention according to the intention concept corresponding to the matched node and the keyword specifically comprises:
searching a target word matched with the keyword from a preset word bank;
Determining the semantics of the keywords according to the semantics of the target words;
and acquiring the intention of the user according to the intention concept corresponding to the matched node and the semantics of the keyword.
4. The method for recognizing the intention based on the concept graph according to any one of claims 1-3, wherein after parsing the voice information and extracting the question words and the keywords in the voice information, the method further comprises:
when the keywords contain specific keywords, receiving a touch signal of the user;
acquiring image information of the user selection area according to the touch signal;
the obtaining the intention of the user according to the intention concept corresponding to the matched node and the keyword specifically includes:
and acquiring the intention of the user according to the intention concept corresponding to the matched node, the keyword and the image information.
5. The method as claimed in claim 4, wherein the obtaining the user's intention according to the intention concept corresponding to the matched node, the keyword and the image information specifically comprises:
Identifying character information in the image information;
replacing the text information with a specific keyword in the keywords;
and acquiring the intention of the user according to the intention concept corresponding to the matched node, the keyword and the text information.
6. An intention recognition apparatus based on a concept graph, comprising:
the acquisition module is used for acquiring voice information of a user;
the extraction module is used for carrying out grammar analysis on the voice information and extracting the question words and the key words in the voice information;
the matching module is used for matching the question words with nodes in a question concept graph to obtain matched nodes, wherein the question concept graph is constructed in advance;
and the intention acquisition module is used for acquiring the intention of the user according to the intention concept corresponding to the matched node and the keyword.
7. The concept graph-based intention recognition apparatus according to claim 6, further comprising a construction module;
the building module comprises:
the corpus collection unit is used for collecting a large amount of user corpuses;
the word extraction unit is used for extracting the question words in the corpus;
The concept acquisition unit is used for acquiring the intention concepts of the same-intention question words in the corpus;
the relationship establishing unit is used for establishing a mapping relationship between the question and legal terms and the intention concepts;
and the construction unit is used for constructing a question and law concept graph according to the mapping relation between the question and law words and the intention concepts.
8. The concept graph-based intention recognition device according to claim 6, wherein the intention acquisition module comprises:
the word searching unit is used for searching a target word matched with the keyword from a preset word bank;
the semantic determining unit is used for determining the semantics of the keywords according to the semantics of the target words;
and the intention acquisition unit is used for acquiring the intention of the user according to the intention concept corresponding to the matched node and the semantics of the keyword.
9. The concept graph-based intention recognition apparatus according to any one of claims 6 to 8, further comprising:
the signal receiving module is used for receiving a touch signal of the user when the keyword contains a specific keyword;
the image acquisition module is used for acquiring the image information of the user selection area according to the touch signal;
The intention acquisition module is further used for acquiring the intention of the user according to the intention concept corresponding to the matched node, the keyword and the image information.
10. The concept graph-based intention recognition device according to claim 9, wherein the intention acquisition module comprises:
the identification unit is used for identifying character information in the image information;
the replacing unit is used for replacing the text information with a specific keyword in the keywords;
and the intention acquisition unit is used for acquiring the intention of the user according to the intention concept corresponding to the matched node, the keyword and the character information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910329359.6A CN111858840A (en) | 2019-04-23 | 2019-04-23 | Intention identification method and device based on concept graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910329359.6A CN111858840A (en) | 2019-04-23 | 2019-04-23 | Intention identification method and device based on concept graph |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111858840A true CN111858840A (en) | 2020-10-30 |
Family
ID=72952187
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910329359.6A Pending CN111858840A (en) | 2019-04-23 | 2019-04-23 | Intention identification method and device based on concept graph |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111858840A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112463920A (en) * | 2020-11-25 | 2021-03-09 | 联想(北京)有限公司 | Information response method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110093479A1 (en) * | 2009-10-19 | 2011-04-21 | Vexigo, Ltd. | System and method for use of semantic understanding in storage, searching and providing of data or other content information |
| CN108228820A (en) * | 2017-12-30 | 2018-06-29 | 厦门太迪智能科技有限公司 | User's query intention understanding method, system and terminal |
| CN109582825A (en) * | 2018-12-07 | 2019-04-05 | 百度在线网络技术（北京）有限公司 | Method and apparatus for generating information |
-
2019
- 2019-04-23 CN CN201910329359.6A patent/CN111858840A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110093479A1 (en) * | 2009-10-19 | 2011-04-21 | Vexigo, Ltd. | System and method for use of semantic understanding in storage, searching and providing of data or other content information |
| CN108228820A (en) * | 2017-12-30 | 2018-06-29 | 厦门太迪智能科技有限公司 | User's query intention understanding method, system and terminal |
| CN109582825A (en) * | 2018-12-07 | 2019-04-05 | 百度在线网络技术（北京）有限公司 | Method and apparatus for generating information |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112463920A (en) * | 2020-11-25 | 2021-03-09 | 联想(北京)有限公司 | Information response method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107291783B (en) | Semantic matching method and intelligent equipment | |
| CN106776711B (en) | Chinese medical knowledge map construction method based on deep learning | |
| US8560301B2 (en) | Apparatus and method for language expression using context and intent awareness | |
| CN106601237B (en) | Interactive voice response system and voice recognition method thereof | |
| CN110727779A (en) | Question-answering method and system based on multi-model fusion | |
| KR102041621B1 (en) | System for providing artificial intelligence based dialogue type corpus analyze service, and building method therefor | |
| CN110457689B (en) | Semantic processing method and related device | |
| CN108287858A (en) | The semantic extracting method and device of natural language | |
| CN107301163B (en) | Formula-containing text semantic parsing method and device | |
| US9015168B2 (en) | Device and method for generating opinion pairs having sentiment orientation based impact relations | |
| JP2006190006A5 (en) | ||
| CN106570180A (en) | Artificial intelligence based voice searching method and device | |
| CN115858758A (en) | Intelligent customer service knowledge graph system with multiple unstructured data identification | |
| KR20190059084A (en) | Natural language question-answering system and learning method | |
| CN109949799B (en) | Semantic parsing method and system | |
| CN111414763A (en) | Semantic disambiguation method, device, equipment and storage device for sign language calculation | |
| CN107436916B (en) | Intelligent answer prompting method and device | |
| CN108710653B (en) | On-demand method, device and system for reading book | |
| CN109918677B (en) | English word semantic analysis method and system | |
| CN109346108B (en) | Operation checking method and system | |
| CN110008314B (en) | Intention analysis method and device | |
| CN105096945A (en) | Voice recognition method and voice recognition device for terminal | |
| CN111858840A (en) | Intention identification method and device based on concept graph | |
| CN114528851B (en) | Reply sentence determination method, reply sentence determination device, electronic equipment and storage medium | |
| CN114492362B (en) | Method and system for generating research and report questions and answers and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |