CN112069301A - Intention recognition method, device, server and storage medium - Google Patents
Intention recognition method, device, server and storage medium Download PDFInfo
- Publication number
- CN112069301A CN112069301A CN202010953728.1A CN202010953728A CN112069301A CN 112069301 A CN112069301 A CN 112069301A CN 202010953728 A CN202010953728 A CN 202010953728A CN 112069301 A CN112069301 A CN 112069301A
- Authority
- CN
- China
- Prior art keywords
- intention
- type
- score
- media resource
- resource
- 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
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
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- 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
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- 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
- G06F16/338—Presentation of query results
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
Abstract
The application discloses an intention identification method, an intention identification device, a server and a storage medium, wherein the intention identification method comprises the following steps: when the intention of a plurality of types of media resources is obtained in a sentence to be identified of a target user, a resource heat score of the intention of each type of media resource in the intentions of the plurality of types of media resources is obtained; acquiring a use habit score corresponding to the intention of each type of media resource based on application data of the application of each type of media resource corresponding to the target user; determining an intention score corresponding to the intention of each type of media resource based on the resource popularity score of the intention of each type of media resource and the use habit score corresponding to the intention of each type of media resource; and acquiring intentions with intention scores meeting preset score conditions from the intentions of the various media resources, and taking the intentions as target intentions corresponding to the sentences to be recognized. The method can determine the final recognition intention by referring to the heat of the media resources and the habit of the user when the intentions of the various media resources are recognized, and improves the accuracy of intention recognition.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to an intention recognition method, an intention recognition apparatus, a server, and a storage medium.
Background
Natural Language Processing (NLP) is an important direction in the fields of computer science and artificial intelligence, and with the development of natural Language Processing, natural Language Processing can be applied to the process of intention recognition, i.e., the process of analyzing and understanding statements input by a user and analyzing the intention of the user so as to meet the needs of the user. In the intention recognition, the accuracy of intention recognition influences the experience of a user, however, the related art often has the problem that the accuracy of intention recognition is low.
Disclosure of Invention
In view of the above, the present application proposes an intention recognition method, apparatus, server, and storage medium.
In a first aspect, an embodiment of the present application provides an intention identification method, where the method includes: when the intention of a plurality of types of media resources is obtained by identifying a sentence to be identified of a target user, obtaining a resource heat score of the intention of each type of media resource in the intentions of the plurality of types of media resources; acquiring a use habit score corresponding to the intention of each type of media resource based on application data of the application of each type of media resource corresponding to the target user; determining an intention score corresponding to the intention of each type of media resource based on the resource popularity score of the intention of each type of media resource and the usage habit score corresponding to the intention of each type of media resource; and acquiring the intention of which the intention score meets a preset score condition from the intentions of the various media resources, and taking the intention as a target intention corresponding to the sentence to be recognized.
In a second aspect, an embodiment of the present application provides an intention identification apparatus, including: the system comprises a first score acquisition module, a second score acquisition module, a third score acquisition module and an intention determination module, wherein the first score acquisition module is used for acquiring a resource hot score of the intention of each media resource in the intentions of multiple media resources when the intention of the media resources of multiple categories is acquired by a sentence to be identified of a target user; the second score obtaining module is used for obtaining a usage habit score corresponding to the intention of each type of media resource based on the application data of the application of each type of media resource corresponding to the target user; the third score obtaining module is used for determining an intention score corresponding to the intention of each type of media resource based on the resource hot score of the intention of each type of media resource and the usage habit score corresponding to the intention of each type of media resource; the intention determining module is used for obtaining the intention of which the intention score meets a preset score condition from the intentions of the various media resources, and the intention is used as the target intention corresponding to the sentence to be recognized.
In a third aspect, an embodiment of the present application provides a server, including: one or more processors; a memory; one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to perform the intent recognition method provided by the first aspect described above.
In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, where a program code is stored in the computer-readable storage medium, and the program code may be called by a processor to execute the intention identification method provided in the first aspect.
The scheme provided by the application can refer to the popularity of each type of media resource and the application habits of the user on each type of media resource when recognizing the intentions of the plurality of types of media resources by acquiring the resource popularity score of the intention of each type of media resource in the intentions of the target user, acquiring the use habit score corresponding to the intention of each type of media resource based on the application data of the application of each type of media resource corresponding to the target user, determining the intention score corresponding to the intention of each type of media resource based on the resource popularity score of each type of media resource and the use habit score corresponding to the intention of each type of media resource, and then acquiring the intention of which the intention score meets the preset score condition from the intentions of the plurality of types of media resources as the target intention corresponding to the sentence to be recognized, and determining a final intention recognition result, and improving the accuracy of intention recognition.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 shows a flow diagram of an intent recognition method according to one embodiment of the present application.
FIG. 2 shows a flowchart of an intent recognition method according to another embodiment of the present application.
Fig. 3 shows a flowchart of step S210 in an intention identifying method according to another embodiment of the present application.
Fig. 4 shows a flowchart of step S220 in an intention identifying method according to another embodiment of the present application.
Fig. 5 shows a flowchart of step S230 in an intention identifying method according to another embodiment of the present application.
FIG. 6 shows a block diagram of an intent recognition apparatus, according to one embodiment of the present application.
Fig. 7 is a block diagram of a server for performing an intention recognition method according to an embodiment of the present application.
Fig. 8 is a storage unit for storing or carrying a program code implementing an intention recognition method according to an embodiment of the present application.
Detailed Description
In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application.
With the development of Automatic Speech Recognition technology (Automatic Speech Recognition), the accuracy rate thereof is continuously improved, and the Speech Recognition technology has been widely appeared in the life of people. In addition, with the development of artificial intelligence technology, natural language processing technology has also developed suddenly and rapidly, and applications of dialog systems or robots based on natural language processing in devices such as mobile phones and televisions are also popular among people.
Natural language processing is an important direction in the fields of computer science and artificial intelligence, and researches various theories and methods for realizing effective communication between people and computers by using natural language. Natural language processing is a science integrating linguistics, computer science and mathematics. Therefore, the research in this field will involve natural language, i.e. the language that people use everyday, so it is closely related to the research of linguistics. Natural language processing techniques typically include text processing, semantic understanding, machine translation, information retrieval, robotic question-and-answer systems, and knowledge-maps.
The natural language processing can be applied to the process of intention identification, namely, the process of analyzing and understanding retrieval information input by a user and analyzing the intention of the user so as to be beneficial to meeting the search requirement of the user. When natural language processing is used in intent recognition, the accuracy of intent recognition will impact the user's experience. In the related art, when a plurality of intentions are recognized based on natural language processing, it is possible to perform processing according to the recognized plurality of intentions or to select one of the intentions for processing according to a rule, but there is a case where it is inaccurate, greatly affecting the user experience. For example, in a media application scenario, when a user says "play no-way", "no-way" is music, movie, etc., but the user often only wants a desired result, usually, the user will force a certain type of intention to be determined according to certain conditions, for example, in sound, the priority of music is higher than that of movie, in television, the priority of movie is higher than that of music. However, since the habits and needs of each user are different, there may be a case where the user needs cannot be satisfied by forcibly determining an intention.
In view of the above problems, the inventors provide an intention identification method, an apparatus, a server, and a storage medium according to embodiments of the present application, which can determine a final intention identification result by referring to the heat of media resources and user habits when identifying intentions of multiple types of media resources, thereby improving accuracy of intention identification. The specific intention recognition method is described in detail in the following examples.
Referring to fig. 1, fig. 1 is a schematic flow chart illustrating an intention identification method according to an embodiment of the present application. In a specific embodiment, the intention recognition method is applied to the intention recognition apparatus 400 shown in fig. 6 and the server 100 (fig. 7) configured with the intention recognition apparatus 400. The specific flow of the embodiment will be described below by taking a server as an example, and it is understood that the server applied in the embodiment may be a conventional server, a cloud server, and the like, and is not limited herein. As will be explained in detail with respect to the flow shown in fig. 1, the intention identifying method may specifically include the following steps:
step S110: when the intention of multiple media resources is obtained by identifying the sentence to be identified of the target user, the resource heat score of the intention of each media resource in the intentions of the multiple media resources is obtained.
In the embodiment of the application, the server can acquire the sentence to be recognized of the target user and perform intention recognition on the sentence to be recognized. The sentence to be recognized can be sent to the server by the target user through the electronic device, and correspondingly, the server can receive the sentence to be recognized sent by the electronic device. The target user may be any user, and is not limited herein.
In some embodiments, the electronic device may collect voice input by a target user, recognize the collected voice by using a voice recognition technology, thereby obtaining a sentence to be recognized, and send the sentence to be recognized to a server; the electronic equipment can also detect the input operation of the target user in the interactive interface so as to detect the sentence to be recognized input by the target user in the interactive interface and send the detected sentence to be recognized input by the target user in the interactive interface to the server; the electronic equipment can also transmit the collected voice to the server when collecting the voice input by the target user, and the server performs voice recognition on the received voice to further obtain the sentence to be recognized. Of course, the manner in which the electronic device obtains the sentence to be recognized may not be limited.
In some embodiments, when the server performs intent recognition on a sentence to be recognized to determine an intent of a target user, the server may perform word segmentation on the sentence to be recognized to obtain a plurality of keywords, and generate a recognition feature vector according to the obtained plurality of keywords; then, the similarity between the feature vector and the feature vectors of various intentions is identified; and finally, screening out the intentions with the similarity greater than the specified similarity as intentions obtained by identifying the sentences to be identified. The recognition characteristic vectors can be word vectors of keywords, the characteristic vectors of various intentions can also be word vectors of the keywords, and the intentions obtained through recognition can be determined through similarity calculation among the word vectors.
In other embodiments, the server may also recognize the intention of the sentence to be recognized through a pre-trained intention recognition model. Specifically, the server may perform word segmentation processing on the sentence to be recognized to obtain a plurality of keywords; then quantizing the plurality of keywords obtained by processing into windowed word vectors; inputting the obtained word vectors into an intention recognition model to obtain the probability of various intentions output by the intention recognition model; and finally, according to the probabilities of various intentions, determining the intentions with the probability greater than the specified probability as the intentions obtained by identifying the basis to be identified. The intention recognition model may be obtained by training an initial model according to word vectors of the keywords corresponding to a large number of sentences and labeled intents, and the initial model may be a convolutional neural network or the like, which is not limited herein.
In still other embodiments, the server may also identify the intent of the sentence to be identified based on a pre-constructed knowledge graph of intents. Specifically, the server may perform word segmentation and entity identification on the sentence to be identified, and obtain a subject, a predicate, an object, and the like in the sentence to be identified according to the syntactic analysis; and then finding out a matching intention from a pre-constructed knowledge graph according to the extracted subject, predicate and object, wherein the matching intention is used as the identified intention.
Of course, in the embodiment of the present application, the specific manner of performing intent recognition on the sentence to be recognized may not be limited.
In the embodiment of the present application, after performing intent recognition on a sentence to be recognized, the server may determine whether the recognized intent includes intentions of multiple types of media resources, that is, intentions of a media application scene, for example, music "a" is played, video "a" is played, broadcast "a" is played, and the like. It is to be understood that the intent of a particular media asset may not be limiting, as it relates to the media asset. Wherein the server may first determine whether the identified intent includes an intent for a media asset and then determine whether there are intents for multiple categories of media assets. Wherein multiple categories of media assets may refer to different categories of media assets, such as music, video, broadcast, etc. When the server determines that the sentence to be recognized is subjected to intention recognition to obtain the intentions of various media resources, the intention of various media resources of the sentence to be recognized of the current target user is represented, and at the moment, an intention needs to be determined to serve as the intention of the target user.
In the embodiment of the application, when determining that the intention of multiple types of media resources is obtained by performing intention identification on a sentence to be identified, the server may obtain a resource hotness score of the intention of each type of resources in the intentions of the multiple types of media resources, and may also obtain a usage habit score corresponding to the intention of each type of media resources, so as to determine the intention score of the intention of each type of media resources, and further determine a final intention.
In some implementations, the server can determine the intent of each type of media asset to be hot in the history identification data as an asset hot score based on the history identification data. For example, the history number of times that the history is identified as the intention of each media resource may be determined based on the final intention identified by the history, and then the popularity score of each media resource may be obtained according to the ratio of the history number to the total number of times of identification in the history identification data. Therefore, the popularity score of the intention of each type of media resource is determined according to the historical identification data, so that the intention identification conditions of different users can be referred when the final intention is determined subsequently, and the accuracy of intention identification is improved.
In other embodiments, the server may also obtain the heat of the media resources required in the above intentions of each type of media resources according to the historical search data. For example, the server may obtain the popularity score of the intent of each type of media resource based on a ratio of the number of searches for the above-identified media resource in the domain of each type of media resource to the total number of searches for the media resource in the domain of that type of media resource. Therefore, the popularity score of the intention of each type of media resource is determined according to the historical search data, so that the intention identification accuracy can be improved by referring to the search conditions of different users for the resources when the final intention is determined.
Of course, the way in which the server specifically obtains the intended resource hotness score of each type of media resource may not be limited.
Step S120: and acquiring a use habit score corresponding to the intention of each type of media resource based on the application data of the application of each type of media resource corresponding to the target user.
In this embodiment of the application, the electronic device may further determine, based on the application data of the application of each type of media resource corresponding to the target user, a usage habit score corresponding to each type of media resource. The application data may be usage data, installation data, and the like of the application of each type of resource by the target user, and the application data reflects the usage habits of the application of each type of media resource by the user, and the data is quantized, so that a usage habit score corresponding to the intention of each type of media resource can be obtained.
Step S130: and determining an intention score corresponding to the intention of each type of media resource based on the resource hot score of the intention of each type of media resource and the use habit score corresponding to the intention of each type of media resource.
In this embodiment of the application, after obtaining the resource popularity score of the intention of each type of media resource and the usage habit score corresponding to the intention of each type of media resource, the server may determine the intention score corresponding to the intention of each type of media resource based on the resource popularity score and the usage habit score.
In some embodiments, the server may perform weighted summation on the resource hot score of the intention of each type of media resource and the usage habit score corresponding to the intention of each type of media resource according to the weight values corresponding to the resource hot score and the usage habit score, respectively, so as to obtain the intention score of each type of media resource.
Step S140: and acquiring the intention of which the intention score meets a preset score condition from the intentions of the various media resources, and taking the intention as a target intention corresponding to the sentence to be recognized.
In the embodiment of the application, after determining the intention score of each type of media resource, the server may determine, according to the intention score of each type of media resource, an intention whose intention score meets a preset score condition from the intentions of the various types of media resources, as a target intention corresponding to the sentence to be recognized, that is, an intention finally obtained by performing intention recognition on the sentence to be recognized.
In some embodiments, the preset scoring condition may include: the intention score is the highest, or the intention score is greater than a preset score, and the specific preset score condition may not be limited. Therefore, the intention scores of the intentions of each type of media resources are determined, the resource heat and the habits of the user are referred, the intentions of the various types of media resources are screened, and more accurate intentions can be determined to meet the requirements of the user. For example, for the sentence to be recognized "play C1", it is recognized that "musical intention: music is played, and the slot position: c1 "," video intention: playing a video, and positioning: c1 "and" broadcast intent: broadcasting, slot position: c1 ", calculating the intention scores of the music intention, the video intention and the broadcast intention respectively, and if the intention score of the video intention meets the preset score condition, then calculating the intention score of the video intention: playing a video, and positioning: c1' is taken as the finally determined target intention.
In some embodiments, if the preset score condition is greater than the preset score value, and the server screens out the intentions of the at least two types of media resources according to the preset score condition, the server may further determine whether the electronic device is installed with applications corresponding to the at least two types of media resources, and screen out the intentions of the at least two types of media resources for installation of corresponding applications according to the determination result.
The intention identification method provided by the embodiment of the application, when the intention of a plurality of types of media resources is obtained by identifying a sentence to be identified of a target user, obtains a resource hot score of the intention of each type of media resource in the intentions of the plurality of types of media resources, obtains a use habit score corresponding to the intention of each type of media resource based on application data of each type of media resource corresponding to the target user, determines the intention score corresponding to the intention of each type of media resource based on the resource hot score of the intention of each type of media resource and the use habit score corresponding to the intention of each type of media resource, then obtains the intention of which score meets a preset score condition from the intention of the plurality of types of media resources as the target intention corresponding to the sentence to be identified, thereby realizing that when the intention of the plurality of types of media resources is identified, the hot degree of each type of media resource and the use habit of the application of each type of media resource by the user are referred, and determining a final intention recognition result, and improving the accuracy of intention recognition.
Referring to fig. 2, fig. 2 is a schematic flow chart illustrating an intention identification method according to another embodiment of the present application. The intention identification method is applied to the server, and will be explained in detail with respect to the flow shown in fig. 2, and specifically may include the following steps:
step S210: when the intention of multiple types of media resources is obtained by recognizing the sentence to be recognized of the target user, a first resource heat score corresponding to the intention of each type of media resources is obtained based on corresponding historical recognition data in the field of each type of media resources.
In the embodiment of the application, when the server identifies the intention of the sentence to be identified of the target user to obtain multiple types of media resources and obtains the resource heat score of each type of media resource, the server can determine the resource heat score according to the historical identification data and the historical viewing data of the result of each type of media resource. The resource popularity score may include a first resource popularity score that may be calculated based on the historical identification data and a second resource popularity score that may be calculated based on the historical viewing data of the results for each type of media resource.
In some embodiments, since the intention of each of the above media resources is recognized according to the same sentence to be recognized, the intention of the media resources includes the same entity name. The entity name may be a name of a media resource, for example, for the sentence "play C1" to be recognized, a "music intention: music is played, and the slot position: c1 "," video intention: playing a video, and positioning: c1 "and" broadcast intent: broadcasting, slot position: c1 ", the same entity name contained is C1.
In some embodiments, referring to fig. 3, step S210 may include:
step S211: acquiring the quantity of historical identification sentences containing the entity names in the fields corresponding to the media resources of each type as a target quantity;
step S212: acquiring the total number of historical identification sentences in the field corresponding to each type of media resource;
step S213: and acquiring the ratio of the target quantity to the total quantity to be used as the first resource heat score.
In this embodiment, when the first resource popularity score is calculated based on the historical identification data, for example, the first resource popularity score is calculated for the music intent, the number a of identification sentences including the entity names and including the music domain keywords and the total number B of identification sentences including the music domain keywords may be obtained, and then the first resource popularity score of the music intent may be obtained according to the ratio of the number a to the total number B. And aiming at the intentions of various media resources, calculating the first resource heat score, namely obtaining the first heat score of the intentions of each media resource.
In the mode, when the target number and the total number are obtained, the historical recognition sentences can be screened according to labels in different fields marked on the historical recognition sentences in the on-line skill, and then the target number and the total number are determined, so that the accuracy is improved.
In other embodiments, the history identification data may be used to determine the history times of the intentions of each type of media resource, and then the popularity score of each type of media resource may be obtained according to the ratio of the history times to the total identification times in the history identification data.
Step S220: and acquiring a second resource hot score corresponding to the intention of each type of media resource based on the history viewing data of the result of each type of media resource in the history intention identification result.
Similarly, since the intention of each media resource is identified according to the same sentence to be identified, the intention of each media resource includes the same entity name.
In some embodiments, referring to fig. 4, step S220 may include:
step S221: acquiring historical viewing data containing a historical intention identification result of the entity name;
step S222: and acquiring the viewing rate of the result of each type of media resource according to the historical viewing data, and taking the viewing rate as a second resource hot score corresponding to the intention of each type of media resource.
Exemplarily, the above entity name is C2, and in the history recognition, the history intention recognition result output according to C2 includes: and correspondingly, outputting corresponding video resources, song resources and novel resources for a user to check, and counting the check rate of the video resources, the song resources and the novel resources according to the check condition of the video resources, the song resources and the novel resources of the user when the intention identification result is output each time. The viewing rate can be determined according to the ratio of the number of times of viewing the result of the intention of various media resources to the total number of times of viewing in the historical viewing data. The above history viewing data may be obtained from a third party platform, such as a resource search platform, and the like.
Step S230: and acquiring a first use habit score corresponding to the intention of each type of media resource based on the use data of the application of each type of media resource corresponding to the target user.
In the embodiment of the present application, the application data may be usage data, installation data, and the like of the application of each type of resource by the target user. The usage habit score corresponding to the intent of each type of media asset may include a first usage habit score and a second usage habit score. The first usage habit score can be calculated according to the usage data of the application of each type of resource by the target user, and the second usage habit score can be calculated according to the installation data of the application of each type of resource.
In some embodiments, referring to fig. 5, step S230 may include:
step S231: normalizing the use duration of the application of each type of media resource corresponding to the target user to obtain a normalized value corresponding to the use duration of the application of each type of media resource;
step S232: and determining a normalization value corresponding to the use duration of the application of each type of media resource as a first use habit score corresponding to the intention of each type of media resource.
In this embodiment, the server may determine the duration of each application in the applications of each media field according to the usage data of the applications of each media field of the target user, and then calculate the usage duration of the applications of each media field; then, the sum of the use duration of the application in each media field is obtained, and the ratio of the use duration of the application in each media field to the sum is used to obtain a normalized value.
For example, as shown in the following table:
in the table, if the duration of use of the application in the music field is 18, the duration of use of the application in the video field is 26, and the duration of use of the application in the radio field is 4, the sum is calculated to be 48, and the ratios of 18, 26, 4, and 48 are calculated, respectively, to obtain a normalized value: 0.375, 0.541, and 0.083, resulting in a first usage habit score for the intent of these three categories of media assets.
In some manners, the server may also obtain a ratio of the usage duration of the application in the field of each type of media resource to a specified value after determining the usage duration of the application in the field of each type of media resource according to the application data of the target user, where the specified value is used to normalize the usage duration of the application in the field of each type of media resource, and a specific value may not be limited.
Step S240: and acquiring a second use habit score corresponding to the intention of each type of media resource based on the installation data of the application of each type of media resource corresponding to the target user.
In some embodiments, when the server obtains the second usage habit score corresponding to the intention of each type of media resource based on the installation data of the application of each type of media resource corresponding to the target user, the server may binarize the installation data of the application of each type of media resource corresponding to the target user to obtain a binarized value of the installation data of the application of each type of media resource, and then determine the binarized value of the installation data of the application of each type of media resource as the second usage habit score corresponding to the intention of each type of media resource.
Illustratively, when the intention of acquiring the music resource and the intention of acquiring the video resource by the sentence to be recognized are recognized, if the electronic device of the target user is provided with the application of the music type, the binary value of the installation data is 1, and if the electronic device of the target user is not provided with the application of the music type, the binary value of the installation data is 0; if the electronic equipment of the target user is provided with the video type application, the binarization value of the installation data is 1; if the electronic device of the target user is installed with the video-type application, the binary value of the installation data is 0.
Step S250: and according to the weight of the resource popularity score and the weight of the use habit score, carrying out weighted calculation on the resource popularity score of the intention of each type of media resource and the use habit score corresponding to the intention of each type of media resource to obtain the intention score corresponding to the intention of each type of media resource.
In this embodiment of the present application, a weight corresponding to each type of media resource may be preset, and the server performs weighted summation on the resource popularity score of the intention of each type of media resource and the usage habit score corresponding to the intention of each type of media resource according to the preset weight, so as to obtain the intention score corresponding to the intention of each type of media resource.
For example, the first usage habit score is a, the second usage habit score is b, the second resource heat score is c, and the first resource heat score is d, then the following formula can be used to calculate:
score＝0.4a+o.1b+o.25c+0.6d
wherein 0.4 is the weight of the first usage habit score, 0.1 is the weight of the second usage habit score, 0.25 is the weight of the second resource popularity score, and 0.6 is the weight of the first resource popularity score.
In some embodiments, when the intention recognition method is used in a scenario of a voice assistant of a mobile terminal, since the intention score of the intention of each type of media resource is low, it is determined that there is no technical recall, that is, there is no determined intention, and there may be a case where a browser is skipped to perform a search. For example, the intent score may be calculated according to the following formula:
score＝0.4a+o.1b+o.25c+0.6d+0.11
wherein, a is the first usage habit score, b is the second usage habit score, c is the second resource heat score, d is the first resource heat score, 0.11 is the bottom score.
Step S260: and acquiring the intention of which the intention score meets a preset score condition from the intentions of the various media resources, and taking the intention as a target intention corresponding to the sentence to be recognized.
In the embodiment of the present application, step S260 may refer to the contents of the foregoing embodiments, which are not described herein again.
Step S270: and pushing the target intention and the media resource corresponding to the target intention to the electronic equipment of the target user.
In the embodiment of the application, after determining the target intention from the intentions of the various media resources, the electronic device may acquire the media resource corresponding to the target intention. For example, if the target intent is "musical intent: music is played, and the slot position: music A', the slot position represents the name of the media resource, and then the music resource corresponding to the music A can be obtained. The server may obtain the media resource locally, or may obtain the media resource from another server. The electronic device may push both the determined target intent, which may be used to indicate the operation of the electronic device, and the media resource corresponding to the target intent, which may be used for the electronic device to present the resource, to the electronic device, for example, the target intent is "music intent: music is played, and the slot position: when music a is available, the electronic device may play music a.
In some embodiments, when the target is intended to show the media resource by using the application, but the electronic device does not install the media application of the type, the server can also push the media application to the electronic device, so that the media resource can be smoothly shown after the electronic device installs the media application.
The intention identification method provided by the embodiment of the application,
referring to fig. 6, a block diagram of an intention recognition apparatus 400 according to an embodiment of the present application is shown. The intention recognition device 400 employs the server described above, and the intention recognition device 400 includes: a first score acquisition module 410, a second score acquisition module 420, a third score acquisition module 430, and an intent determination module 440. The first score obtaining module 410 is configured to obtain a resource popularity score of an intention of each category of media resources in the intentions of multiple categories of media resources when an intention of multiple categories of media resources is obtained in a sentence to be recognized of a target user; the second score obtaining module 420 is configured to obtain a usage habit score corresponding to an intention of each type of media resource based on application data of an application of each type of media resource corresponding to the target user; the third score obtaining module 430 is configured to determine an intention score corresponding to the intention of each type of media resource based on the resource hotness score of the intention of each type of media resource and the usage habit score corresponding to the intention of each type of media resource; the intention determining module 440 is configured to obtain, from the intentions of the multiple categories of media resources, an intention that the intention score satisfies a preset score condition, as a target intention corresponding to the sentence to be recognized.
In some embodiments, the application data comprises usage data for an application and installation data for the application, the usage habit score comprising a first usage habit score and a second usage habit score. The second score obtaining module 420 may include: the device comprises a first acquisition unit and a second acquisition unit. The first obtaining unit is used for obtaining a first usage habit score corresponding to the intention of each type of media resource based on the usage data of the application of each type of media resource corresponding to the target user; the second obtaining unit is used for obtaining a second usage habit score corresponding to the intention of each type of media resource based on the installation data of the application of each type of media resource corresponding to the target user.
In this embodiment, the usage data includes a length of time of use. The first obtaining unit may specifically be configured to: normalizing the use duration of the application of each type of media resource corresponding to the target user to obtain a normalized value corresponding to the use duration of the application of each type of media resource; and determining a normalization value corresponding to the use duration of the application of each type of media resource as a first use habit score corresponding to the intention of each type of media resource.
In this embodiment, the second obtaining unit may specifically be configured to: binarizing the installation data of the application of each type of media resource corresponding to the target user to obtain a binarized value of the installation data of the application of each type of media resource; and determining the binary value of the application installation data of each type of media resource as a second usage habit score corresponding to the intention of each type of media resource.
In some embodiments, the resource popularity score includes a first resource popularity score and a second resource popularity score. The first score obtaining module 410 may include: a third acquisition unit and a fourth acquisition unit. The third acquisition unit is used for acquiring a first resource hot score corresponding to the intention of each type of media resource based on corresponding historical identification data in the field of each type of media resource; the fourth obtaining unit is used for obtaining a second resource hot score corresponding to the intention of each type of media resource based on the history viewing data of the result of each type of media resource in the history intention identification result.
In this embodiment, the intent of each type of media asset contains the same entity name. The third obtaining unit may specifically be configured to: acquiring the quantity of historical identification sentences containing the entity names in the fields corresponding to the media resources of each type as a target quantity; acquiring the total number of historical identification sentences in the field corresponding to each type of media resource; and acquiring the ratio of the target quantity to the total quantity to be used as the first resource heat score.
In this embodiment, the intent of each type of media asset contains the same entity name. The fourth obtaining unit may specifically be configured to: acquiring historical viewing data containing a historical intention identification result of the entity name; and acquiring the viewing rate of the result of each type of media resource according to the historical viewing data, and taking the viewing rate as a second resource hot score corresponding to the intention of each type of media resource.
In some embodiments, the third score acquisition module 430 may be specifically configured to: and according to the weight of the resource popularity score and the weight of the use habit score, carrying out weighted calculation on the resource popularity score of the intention of each type of media resource and the use habit score corresponding to the intention of each type of media resource to obtain the intention score corresponding to the intention of each type of media resource.
In some embodiments, the intention recognition apparatus 400 may further include: and an information pushing module. The information pushing module is configured to, after the intent determining module 440 obtains the intent, of which the intent score meets a preset score condition, from the intentions of the multiple categories of media resources, and uses the intent as a target intent corresponding to the sentence to be recognized, push the target intent and the media resource corresponding to the target intent to the electronic device of the target user.
It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described apparatuses and modules may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the several embodiments provided in the present application, the coupling between the modules may be electrical, mechanical or other type of coupling.
In addition, functional modules in the embodiments of the present application may be integrated into one processing module, or each of the modules may exist alone physically, or two or more modules are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode.
In summary, the solution provided by the present application, when recognizing the intention of a sentence to be recognized of a target user to obtain multiple categories of media resources, obtains a resource popularity score of the intention of each category of media resources from the intentions of the multiple categories of media resources, obtains a usage habit score corresponding to the intention of each category of media resources based on application data of an application of each category of media resources corresponding to the target user, determines an intention score corresponding to the intention of each category of media resources based on the resource popularity score of the intention of each category of media resources and the usage habit score corresponding to the intention of each category of media resources, and then obtains an intention whose intention score satisfies a preset score condition from the intentions of the multiple categories of media resources as a target intention corresponding to the sentence to be recognized, so that it is possible to refer to the popularity of each category of media resources and the usage habit of the user to each category of media resources when recognizing the intention of the multiple categories of media resources, and determining a final intention recognition result, and improving the accuracy of intention recognition.
Referring to fig. 7, a block diagram of a server according to an embodiment of the present disclosure is shown. The server 100 may be a conventional server, a cloud server, or the like capable of running an application. The server 100 in the present application may include one or more of the following components: a processor 110, a memory 120, and one or more applications, wherein the one or more applications may be stored in the memory 120 and configured to be executed by the one or more processors 110, the one or more programs configured to perform a method as described in the aforementioned method embodiments.
Processor 110 may include one or more processing cores. The processor 110 connects various parts within the overall server 100 using various interfaces and lines, performs various functions of the server 100 and processes data by executing or executing instructions, programs, code sets, or instruction sets stored in the memory 120, and calling data stored in the memory 120. Alternatively, the processor 110 may be implemented in hardware using at least one of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). The processor 110 may integrate one or more of a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a modem, and the like. Wherein, the CPU mainly processes an operating system, a user interface, an application program and the like; the GPU is used for rendering and drawing display content; the modem is used to handle wireless communications. It is understood that the modem may not be integrated into the processor 110, but may be implemented by a communication chip.
The Memory 120 may include a Random Access Memory (RAM) or a Read-Only Memory (Read-Only Memory). The memory 120 may be used to store instructions, programs, code sets, or instruction sets. The memory 120 may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing various method embodiments described below, and the like. The storage data area may also store data created by the server 100 in use (such as phone books, audio and video data, chat log data), and the like.
Referring to fig. 8, a block diagram of a computer-readable storage medium according to an embodiment of the present application is shown. The computer-readable medium 800 has stored therein a program code that can be called by a processor to execute the method described in the above-described method embodiments.
The computer-readable storage medium 800 may be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read only memory), an EPROM, a hard disk, or a ROM. Alternatively, the computer-readable storage medium 800 includes a non-volatile computer-readable storage medium. The computer readable storage medium 800 has storage space for program code 810 to perform any of the method steps of the method described above. The program code can be read from or written to one or more computer program products. The program code 810 may be compressed, for example, in a suitable form.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not necessarily depart from the spirit and scope of the corresponding technical solutions in the embodiments of the present application.
Claims (12)
1. An intent recognition method, the method comprising:
when the intention of a plurality of types of media resources is obtained by identifying a sentence to be identified of a target user, obtaining a resource heat score of the intention of each type of media resource in the intentions of the plurality of types of media resources;
acquiring a use habit score corresponding to the intention of each type of media resource based on application data of the application of each type of media resource corresponding to the target user;
determining an intention score corresponding to the intention of each type of media resource based on the resource popularity score of the intention of each type of media resource and the usage habit score corresponding to the intention of each type of media resource;
and acquiring the intention of which the intention score meets a preset score condition from the intentions of the various media resources, and taking the intention as a target intention corresponding to the sentence to be recognized.
2. The method according to claim 1, wherein the application data includes usage data of an application and installation data of the application, the usage habit score includes a first usage habit score and a second usage habit score, and the obtaining of the usage habit score corresponding to the intention of each type of media resource based on the application data of the application of each type of media resource corresponding to the target user includes:
acquiring a first use habit score corresponding to the intention of each type of media resource based on the use data of the application of each type of media resource corresponding to the target user;
and acquiring a second use habit score corresponding to the intention of each type of media resource based on the installation data of the application of each type of media resource corresponding to the target user.
3. The method according to claim 2, wherein the usage data includes a usage duration, and the obtaining a first usage habit score corresponding to the intention of each type of media resource based on the usage data of the application of each type of media resource corresponding to the target user comprises:
normalizing the use duration of the application of each type of media resource corresponding to the target user to obtain a normalized value corresponding to the use duration of the application of each type of media resource;
and determining a normalization value corresponding to the use duration of the application of each type of media resource as a first use habit score corresponding to the intention of each type of media resource.
4. The method according to claim 2, wherein the obtaining a second usage habit score corresponding to the intention of each type of media resource based on the installation data of the application of each type of media resource corresponding to the target user comprises:
binarizing the installation data of the application of each type of media resource corresponding to the target user to obtain a binarized value of the installation data of the application of each type of media resource;
and determining the binary value of the application installation data of each type of media resource as a second usage habit score corresponding to the intention of each type of media resource.
5. The method of claim 1, wherein the resource popularity score comprises a first resource popularity score and a second resource popularity score, and wherein the obtaining the resource popularity score for the intent of each of the plurality of categories of media resources comprises:
acquiring a first resource heat score corresponding to the intention of each type of media resource based on corresponding historical identification data in the field of each type of media resource;
and acquiring a second resource hot score corresponding to the intention of each type of media resource based on the history viewing data of the result of each type of media resource in the history intention identification result.
6. The method of claim 5, wherein the intentions of each type of media resource include the same entity name, and the obtaining a first resource hotness score corresponding to the intention of each type of media resource based on the corresponding historical identification data in the domain of each type of media resource comprises:
acquiring the quantity of historical identification sentences containing the entity names in the fields corresponding to the media resources of each type as a target quantity;
acquiring the total number of historical identification sentences in the field corresponding to each type of media resource;
and acquiring the ratio of the target quantity to the total quantity to be used as the first resource heat score.
7. The method of claim 5, wherein the intentions of each type of media resource contain the same entity name, and the obtaining a second resource hotness score corresponding to the intention of each type of media resource based on historical viewing data of the result of each type of media resource in the historical intention identification result comprises:
acquiring historical viewing data containing a historical intention identification result of the entity name;
and acquiring the viewing rate of the result of each type of media resource according to the historical viewing data, and taking the viewing rate as a second resource hot score corresponding to the intention of each type of media resource.
8. The method according to any one of claims 1-7, wherein the determining an intention score corresponding to the intention of each type of media resource based on the resource hot score of the intention of each type of media resource and the usage habit score corresponding to the intention of each type of media resource comprises:
and according to the weight of the resource popularity score and the weight of the use habit score, carrying out weighted calculation on the resource popularity score of the intention of each type of media resource and the use habit score corresponding to the intention of each type of media resource to obtain the intention score corresponding to the intention of each type of media resource.
9. The method according to any one of claims 1 to 7, wherein after obtaining, from the intentions of the multiple categories of media resources, the intention that the intention score satisfies a preset score condition as a target intention corresponding to the sentence to be recognized, the method further comprises:
and pushing the target intention and the media resource corresponding to the target intention to the electronic equipment of the target user.
10. An intent recognition apparatus, characterized in that the apparatus comprises: a first score acquisition module, a second score acquisition module, a third score acquisition module, and an intent determination module, wherein,
the first score acquisition module is used for acquiring the resource hot score of the intention of each media resource in the intentions of the multiple media resources when the intention of the multiple media resources is acquired by a sentence to be identified of a target user;
the second score obtaining module is used for obtaining a usage habit score corresponding to the intention of each type of media resource based on the application data of the application of each type of media resource corresponding to the target user;
the third score obtaining module is used for determining an intention score corresponding to the intention of each type of media resource based on the resource hot score of the intention of each type of media resource and the usage habit score corresponding to the intention of each type of media resource;
the intention determining module is used for obtaining the intention of which the intention score meets a preset score condition from the intentions of the various media resources, and the intention is used as the target intention corresponding to the sentence to be recognized.
11. A server, comprising:
one or more processors;
a memory;
one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to perform the method of any of claims 1-9.
12. A computer-readable storage medium, having stored thereon program code that can be invoked by a processor to perform the method according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010953728.1A CN112069301A (en) | 2020-09-11 | 2020-09-11 | Intention recognition method, device, server and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010953728.1A CN112069301A (en) | 2020-09-11 | 2020-09-11 | Intention recognition method, device, server and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112069301A true CN112069301A (en) | 2020-12-11 |
Family
ID=73696230
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010953728.1A Pending CN112069301A (en) | 2020-09-11 | 2020-09-11 | Intention recognition method, device, server and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112069301A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102955798A (en) * | 2011-08-25 | 2013-03-06 | 腾讯科技（深圳）有限公司 | Search engine based search method and search server |
| CN110784768A (en) * | 2019-10-17 | 2020-02-11 | 珠海格力电器股份有限公司 | Multimedia resource playing method, storage medium and electronic equipment |
| CN111488443A (en) * | 2020-04-08 | 2020-08-04 | 苏州思必驰信息科技有限公司 | Skill selection method and device |
-
2020
- 2020-09-11 CN CN202010953728.1A patent/CN112069301A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102955798A (en) * | 2011-08-25 | 2013-03-06 | 腾讯科技（深圳）有限公司 | Search engine based search method and search server |
| CN110784768A (en) * | 2019-10-17 | 2020-02-11 | 珠海格力电器股份有限公司 | Multimedia resource playing method, storage medium and electronic equipment |
| CN111488443A (en) * | 2020-04-08 | 2020-08-04 | 苏州思必驰信息科技有限公司 | Skill selection method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106897428B (en) | Text classification feature extraction method and text classification method and device | |
| CN108491433B (en) | Chat response method, electronic device and storage medium | |
| CN110069709B (en) | Intention recognition method, device, computer readable medium and electronic equipment | |
| CN111428010B (en) | Man-machine intelligent question-answering method and device | |
| CN111144952A (en) | Advertisement recommendation method, device, server and storage medium based on user interests | |
| CN111985249A (en) | Semantic analysis method and device, computer-readable storage medium and electronic equipment | |
| CN110414004A (en) | A kind of method and system that core information extracts | |
| CN114357973A (en) | Intention recognition method and device, electronic equipment and storage medium | |
| CN113254777B (en) | Information recommendation method and device, electronic equipment and storage medium | |
| CN111312233A (en) | Voice data identification method, device and system | |
| CN111611358A (en) | Information interaction method and device, electronic equipment and storage medium | |
| CN110399473B (en) | Method and device for determining answers to user questions | |
| CN116150306A (en) | Training method of question-answering robot, question-answering method and device | |
| CN111291551A (en) | Text processing method and device, electronic equipment and computer readable storage medium | |
| CN114298007A (en) | Text similarity determination method, device, equipment and medium | |
| CN112069301A (en) | Intention recognition method, device, server and storage medium | |
| CN111858966B (en) | Knowledge graph updating method and device, terminal equipment and readable storage medium | |
| CN115098794A (en) | Public opinion manufacturing group identification method, device, equipment and storage medium | |
| CN115618873A (en) | Data processing method and device, computer equipment and storage medium | |
| CN113673237A (en) | Model training method, intent recognition method, device, electronic equipment and storage medium | |
| CN113434695A (en) | Financial event extraction method and device, electronic equipment and storage medium | |
| CN113505292A (en) | Information pushing method and device, electronic equipment and storage medium | |
| CN114139031B (en) | Data classification method, device, electronic equipment and storage medium | |
| CN115878849B (en) | Video tag association method and device and electronic equipment | |
| CN116383416A (en) | Method and device for requesting multimedia resources |
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 |