CN111966822A - Method and device for determining emotion category of evaluation information - Google Patents
Method and device for determining emotion category of evaluation information Download PDFInfo
- Publication number
- CN111966822A CN111966822A CN201910417377.XA CN201910417377A CN111966822A CN 111966822 A CN111966822 A CN 111966822A CN 201910417377 A CN201910417377 A CN 201910417377A CN 111966822 A CN111966822 A CN 111966822A
- Authority
- CN
- China
- Prior art keywords
- evaluation
- information
- word
- target
- evaluation information
- 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
- 238000011156 evaluation Methods 0.000 title claims abstract description 498
- 230000008451 emotion Effects 0.000 title claims abstract description 177
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000013145 classification model Methods 0.000 claims abstract description 39
- 230000006870 function Effects 0.000 claims description 112
- 230000015654 memory Effects 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 238000013507 mapping Methods 0.000 claims description 6
- 238000012546 transfer Methods 0.000 claims description 5
- 238000010586 diagram Methods 0.000 description 14
- 230000002996 emotional effect Effects 0.000 description 8
- 238000004891 communication Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 230000003287 optical effect Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 230000006403 short-term memory Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- RXKJFZQQPQGTFL-UHFFFAOYSA-N dihydroxyacetone Chemical compound OCC(=O)CO RXKJFZQQPQGTFL-UHFFFAOYSA-N 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000007704 transition 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
- G06F16/35—Clustering; Classification
- G06F16/355—Class or cluster creation or modification
-
- 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
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the application discloses a method and a device for determining emotion categories of evaluation information. One embodiment of the method comprises: and calculating the probability that the words contained in the evaluation information form the evaluation words aiming at the target evaluation object based on a preset evaluation keyword library of the target evaluation object, and carrying out emotion classification on the evaluation information by adopting a trained emotion classification model based on the probability that the words contained in the evaluation information form the evaluation words aiming at the target evaluation object to obtain emotion category information corresponding to the evaluation information and aiming at the target evaluation object. The method and the device can distinguish the emotion types of different evaluation objects represented by the evaluation information, and improve the accuracy of the emotion classification result of the evaluation information.
Description
Technical Field
The embodiment of the application relates to the technical field of computers, in particular to the field of information processing, and particularly relates to a method and a device for determining emotion categories of evaluation information.
Background
With the development of internet technology, more and more internet data is generated. By mining and analyzing the internet data, the potentially valuable information can be accurately acquired, such as habits and preferences of users accessing the internet, and event trends can be predicted.
The evaluation information is information generated by evaluating an event, an acquired article, and the like that an internet user has participated in, and is important internet data. The evaluation information may include objective description of the evaluation object by the user, such as size, color, appearance, etc.; personal emotional tendencies, such as like degrees, for the evaluation subjects may also be included. The emotional tendency of the user in the evaluation information is analyzed, and the evaluation object can be adjusted to adapt to the requirement and preference of the user, so that the reliability of the internet service is improved.
Disclosure of Invention
The embodiment of the application provides a method, a device, electronic equipment and a computer readable medium for determining emotion categories of evaluation information.
In a first aspect, an embodiment of the present disclosure provides a method for determining an emotion category of evaluation information, including: calculating the probability that the words contained in the evaluation information form the evaluation words aiming at the target evaluation object based on a preset evaluation keyword library of the target evaluation object; and performing emotion classification on the evaluation information by adopting a trained emotion classification model based on the probability that the words contained in the evaluation information form an evaluation word aiming at the target evaluation object to obtain emotion category information corresponding to the evaluation information and aiming at the target evaluation object.
In some embodiments, the calculating, based on the preset evaluation keyword library of the target evaluation object, a probability that a word included in the evaluation information constitutes an evaluation word for the target evaluation object includes: splitting the evaluation information into word sequences; constructing a feature function set of a conditional random field based on a preset evaluation keyword library of a target evaluation object, and constructing a conditional probability function based on the feature function set of the conditional random field, wherein the conditional probability function represents the probability of outputting a corresponding evaluation word tag sequence when a text word sequence is input in the conditional random field, and an evaluation word tag in the evaluation word tag sequence is used for identifying that a corresponding input text word forms an evaluation word aiming at the target evaluation object or does not form the evaluation word aiming at the target evaluation object; and calculating the probability that each word in the word sequence forms the appraisal words aiming at the target appraisal object based on the conditional probability function.
In some embodiments, the feature function set includes a first feature function set and a second feature function set, where the first feature function in the first feature function set is used to characterize transfer features between adjacent evaluator labels corresponding to adjacent text words in the input text word sequence, and the second feature function in the second feature function set is used to characterize mapping features of words in the input text word sequence mapped to corresponding evaluator labels; the above feature function set construction conditional probability function based on the conditional random field includes: a conditional probability function is constructed based on the weighted sum of the first feature functions and the weighted sum of the second feature functions.
In some embodiments, the obtaining emotion classification information for the target evaluation object corresponding to the evaluation information by performing emotion classification on the evaluation information using a trained emotion classification model based on the probability that the word included in the evaluation information constitutes the evaluation word for the target evaluation object includes: and converting words in the evaluation information into word vectors, multiplying the word vectors by the corresponding probability of forming the evaluation words aiming at the target evaluation object, inputting the trained emotion classification model for classification, and obtaining emotion category information aiming at the target evaluation object corresponding to the evaluation information.
In some embodiments, the trained emotion classification model includes a long-short term memory network and a classifier; the method further comprises the following steps: and training based on a sample evaluation information set to obtain a trained emotion classification model, wherein the sample evaluation information set comprises sample evaluation information and marking information of the sample evaluation information aiming at the emotion type of the target evaluation object.
In some embodiments, the above method further comprises: and determining that the emotion category information for the target evaluation object is evaluation information of preset emotion category information, serving as to-be-pushed evaluation information, and pushing the to-be-pushed evaluation information to a content providing main body associated with the target evaluation object in the to-be-pushed evaluation information.
In a second aspect, an embodiment of the present disclosure provides an apparatus for determining an emotion category of evaluation information, including: a calculation unit configured to calculate a probability that a word included in the evaluation information constitutes an evaluation word for the target evaluation object based on a preset evaluation keyword library of the target evaluation object; and the classification unit is configured to perform emotion classification on the evaluation information by adopting a trained emotion classification model based on the probability that the words contained in the evaluation information form the evaluation words aiming at the target evaluation object, and obtain emotion category information corresponding to the evaluation information aiming at the target evaluation object.
In some embodiments, the above-mentioned calculation unit is further configured to calculate the probability that the word included in the evaluation information constitutes an evaluation word for the target evaluation object as follows: splitting the evaluation information into word sequences; constructing a feature function set of a conditional random field based on a preset evaluation keyword library of a target evaluation object, and constructing a conditional probability function based on the feature function set of the conditional random field, wherein the conditional probability function represents the probability of outputting a corresponding evaluation word tag sequence when a text word sequence is input in the conditional random field, and an evaluation word tag in the evaluation word tag sequence is used for identifying that a corresponding input text word forms an evaluation word aiming at the target evaluation object or does not form the evaluation word aiming at the target evaluation object; and calculating the probability that each word in the word sequence forms the appraisal words aiming at the target appraisal object based on the conditional probability function.
In some embodiments, the feature function set includes a first feature function set and a second feature function set, where the first feature function in the first feature function set is used to characterize transfer features between adjacent evaluator labels corresponding to adjacent text words in the input text word sequence, and the second feature function in the second feature function set is used to characterize mapping features of words in the input text word sequence mapped to corresponding evaluator labels; the above-mentioned computing unit is further configured to construct the conditional probability function as follows: a conditional probability function is constructed based on the weighted sum of the first feature functions and the weighted sum of the second feature functions.
In some embodiments, the classification unit is further configured to perform emotion classification on the evaluation information as follows, and obtain emotion category information corresponding to the evaluation information and aiming at the target evaluation object: and converting words in the evaluation information into word vectors, multiplying the word vectors by the corresponding probability of forming the evaluation words aiming at the target evaluation object, inputting the trained emotion classification model for classification, and obtaining emotion category information aiming at the target evaluation object corresponding to the evaluation information.
In some embodiments, the trained emotion classification model includes a long-short term memory network and a classifier; the above-mentioned device still includes: and the training unit is configured to train to obtain a trained emotion classification model based on a sample evaluation information set, wherein the sample evaluation information set comprises sample evaluation information and marking information of the sample evaluation information aiming at the emotion category of the target evaluation object.
In some embodiments, the above apparatus further comprises: the pushing unit is configured to determine that the emotion category information of the target evaluation object is evaluation information of preset emotion category information, the evaluation information serves as evaluation information to be pushed, and the evaluation information to be pushed is pushed to a content providing main body related to the target evaluation object in the evaluation information to be pushed.
In a third aspect, an embodiment of the present disclosure provides an electronic device, including: one or more processors; a storage device for storing one or more programs which, when executed by one or more processors, cause the one or more processors to implement the method for determining an emotion classification for rating information as provided in the first aspect.
In a fourth aspect, an embodiment of the present disclosure provides a computer readable medium on which a computer program is stored, wherein the program, when executed by a processor, implements the method for determining an emotion category of rating information provided in the first aspect.
According to the method and the device for determining the emotion category of the evaluation information, the probability that the words contained in the evaluation information form the evaluation words aiming at the target evaluation object is calculated through the preset evaluation keyword library based on the target evaluation object; based on the probability that the words contained in the evaluation information form the evaluation words for the target evaluation object, the trained emotion classification model is adopted to carry out emotion classification on the evaluation information to obtain emotion category information corresponding to the evaluation information and aiming at the target evaluation object, so that emotion classification for different evaluation objects can be distinguished when the evaluation information is subjected to emotion classification, and a more accurate evaluation information emotion classification result can be obtained.
Drawings
Other features, objects and advantages of the present application will become more apparent upon reading of the following detailed description of non-limiting embodiments thereof, made with reference to the accompanying drawings in which:
FIG. 1 is an exemplary system architecture diagram in which embodiments of the present disclosure may be applied;
FIG. 2 is a flow diagram of one embodiment of a method for determining sentiment categories for ratings information according to the present disclosure;
FIG. 3 is a schematic diagram of the computational logic of the long short term memory network in the sentiment classification model of the present disclosure;
FIG. 4 is a schematic diagram of an implementation of a method for determining sentiment categories for rating information according to the present disclosure;
FIG. 5 is a flow diagram of another embodiment of a method for determining sentiment categories for ratings information according to the present disclosure;
FIG. 6 is a schematic structural diagram illustrating an embodiment of an apparatus for determining emotion classifications of ratings information according to the present disclosure;
FIG. 7 is a schematic block diagram of a computer system suitable for use with an electronic device implementing embodiments of the present disclosure.
Detailed Description
The present application will be described in further detail with reference to the following drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant invention and not restrictive of the invention. It should be noted that, for convenience of description, only the portions related to the related invention are shown in the drawings.
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
Fig. 1 shows an exemplary system architecture to which the method for determining an emotion classification of rating information or the apparatus for determining an emotion classification of rating information of the present application may be applied.
As shown in fig. 1, the system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the terminal devices 101, 102, 103 and the server 105. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.
The user 110 may use the terminal devices 101, 102, 103 to interact with the server 105 via the network 104 to receive or send messages or the like. Various life service applications, such as a multimedia playing application, an online shopping application, a search application, a knowledge sharing application, etc., may be installed on the terminal devices 101, 102, 103.
The terminal devices 101, 102, 103 may be various electronic devices that support internet access including, but not limited to, smart phones, tablets, smart watches, electronic books, desktop computers, notebook computers, and the like.
The server 105 may be a server that provides support for the lifestyle service type applications of the terminal device. In the application scenario of the present disclosure, the user 110 may send a service content acquisition request through the terminal devices 101, 102, and 103 to acquire the service content provided by the server 105, and report the service content to the server 105 through the terminal devices 101, 102, and 103 after evaluating the service content. The server 105 may analyze the service content acquisition request sent by the terminal device 101, 102, 103, generate corresponding service content according to the user requirement, and feed back the service content to the terminal device 101, 102, 103. The server 105 may also perform emotion analysis on the evaluation information reported by the terminal devices 101, 102, and 103, and determine an emotion category of the user's evaluation on the service content.
It should be noted that the method for determining the emotion category of evaluation information provided by the embodiment of the present disclosure is generally executed by the server 105, and accordingly, the apparatus for determining the emotion category of evaluation information is generally disposed in the terminal device 101, 102, 103 or the server 105.
It should be understood that the number of terminal devices, networks, and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.
With continued reference to FIG. 2, a flow 200 of one embodiment of a method for determining sentiment categories of rating information according to the present application is shown. The method for determining the emotion category of the evaluation information comprises the following steps:
An executive (e.g., server 105 shown in fig. 1) of the method for determining emotion classifications of ratings information may first obtain the ratings information of the user. Here, the evaluation information is text information generated by evaluating at least one evaluation target by the user.
As an example, the evaluation information may be evaluation information for goods purchased on the e-commerce platform and logistics service, such as "this good quality is good but logistics is slow, or" east-west quality is not so much of a failure but a little brother gives power ".
The evaluation information typically contains relevant information for identifying the targeted evaluation object, such as the type, name, etc. of the evaluation object. The evaluation object can be identified based on a preset evaluation object keyword table. Here, the evaluation target keyword table may include a correspondence table between each evaluation target and a keyword representing the evaluation target. For example, the keyword table for evaluation object includes keywords "logistics", "express delivery", "guo", "delivery", "transportation", and the like corresponding to the evaluation object "logistics", and keywords "goods", "things", "baby", and the like corresponding to the evaluation object "article". In the evaluation information example of "this commodity is good in quality but the physical distribution is somewhat slow", the evaluation object includes "article" and "physical distribution". In the evaluation information example of "something is not so much in quality but is given a lot of force by the courier", the evaluation target also includes "article" and "physical distribution".
In this embodiment, a target evaluation object may be specified in advance, and the emotion category implied by the evaluation information for the specified target evaluation object may be analyzed. For example, the target evaluation object may be designated as "physical distribution". The obtained evaluation information may also include evaluation contents for other evaluation objects, and the emotional tendency of the same piece of evaluation information for different evaluation objects may be different, for example, in the above example of "this good quality, but the logistics is slow", the emotional tendency of the evaluation for "goods" is positive, and the emotional tendency of the evaluation for "logistics" is negative.
In order to avoid the inaccuracy of the emotion types of the evaluation information due to the mutual influence between the emotional tendencies of different evaluation objects in the evaluation information, the emotion types of the target evaluation objects specified in the evaluation information can be analyzed independently.
In this embodiment, an evaluation keyword library of a target evaluation object may be acquired in advance, and the evaluation keyword library of the target evaluation object includes keywords constituting an evaluation word for the target evaluation object. For example, the evaluation keyword library of the target evaluation object "logistics" may include: the method has the advantages of high logistics speed, low delivery speed, good courier attitude, high delivery speed, good package integrity, good brother attitude and the like.
The probability that a word included in the evaluation information constitutes an evaluation word for the target evaluation object may be calculated based on the evaluation keyword library of the target evaluation object. Specifically, the probability that a word included in the evaluation information constitutes an evaluation word for the target evaluation object may be determined based on the degree of matching or the degree of association between the word and an evaluation keyword in the evaluation keyword library of the target evaluation object. For example, if the matching degree between the term included in the evaluation information and the evaluation keyword in the evaluation keyword library reaches a preset threshold interval, it may be determined that the probability that the term constitutes the evaluation word for the target evaluation object is a preset probability value corresponding to the preset threshold interval.
In some alternative implementations, the probability that a word included in the evaluation information constitutes an evaluation word for the target evaluation object may be calculated based on the conditional random field. Specifically, a relational model between the evaluation information and the probability that a word included in the evaluation information constitutes an evaluation word for a target evaluation object may be constructed using a conditional random field, and the probability that a word included in the acquired evaluation information constitutes an evaluation word for a target evaluation object may be calculated using a model based on a conditional random field.
Specifically, the evaluation information may be divided into word sequences by a method such as word segmentation, the word sequences are used as observation sequences of the conditional random field, evaluation term tag sequences formed by tags corresponding to the word sequences and used for characterizing whether each word constitutes an evaluation term for the target evaluation object are used as state sequences of the conditional random field, and a conditional probability model from the observation sequences to the state sequences is constructed.
A feature function set of the conditional random field can be constructed based on a preset evaluation keyword library of the target evaluation object, and then a conditional probability function is constructed based on the feature function set of the conditional random field. The conditional probability function represents the probability of outputting a corresponding evaluation word tag sequence when a text word sequence is input in the conditional random field, and an evaluation word tag in the evaluation word tag sequence is used for identifying whether the corresponding input text word forms an evaluation word aiming at a target evaluation object or does not form the evaluation word aiming at the target evaluation object. For example, when a text word is input to form an evaluation word of a target evaluation object, the corresponding evaluation word label is 1; when the input text word does not constitute the evaluation word of the target evaluation object, the corresponding evaluation word label is 0.
The feature function set of the conditional random field is a feature function for representing an implicit relationship between evaluator tags in an evaluator tag sequence corresponding to the input word sequence. The feature function set of the conditional random field can be designed according to the evaluation keyword library of the target evaluation object.
As an example, one characteristic function may be represented as fi(x，i，yi，yi-1) I ═ 1,2,3, …, n; n is the total number of words in the input word sequence, x is the input word sequence, yiFor the comment label corresponding to the ith input word, yi-1And the evaluation word label corresponding to the i-1 th input word is input. Here, if an evaluator for the target evaluation object is configured, the corresponding evaluator label is 1, otherwise, the corresponding evaluator label is 0. For example, when the (i-1) th word and the (i) th word are both words in the evaluation keyword library, or the (i-1) th word and the (i) th word are combined to form words in the evaluation keyword library, for example, the (i-1) th word is express and the (i) th word is fast, the corresponding characteristic function fi(x，i，yi，yi-1) Can take the value of 1; when the (i-1) th word and the (i) th word are words not in the evaluation keyword library of the target evaluation object, such as 'commodity' and 'good quality', respectively, the corresponding characteristic function fi(x，i，yi，yi-1) May take the value 0.
The method can be adopted to set a feature function set, and construct a conditional probability function P (y | x) based on the feature function set, where x is an input word sequence and y is a corresponding evaluation term tag sequence. For example, all feature functions may be weighted and summed, and then logarithms may be taken as conditional probability functions.
Optionally, the feature function may include a first feature function set and a second feature function set. The first feature function in the first feature function set is used for representing transfer features between adjacent appraiser labels corresponding to adjacent text words in the input text word sequence, and the second feature function in the second feature function set is used for representing mapping features of words in the input text word sequence mapped to the corresponding appraiser labels. The conditional probability function can then be constructed on the basis of the weighted sum of the first characteristic functions and the weighted sum of the second characteristic functions.
The transition characteristics indicate whether the comment labels of adjacent words are transitioned due to certain characteristics, and the mapping characteristics are used for representing whether the words have the characteristics of the corresponding comment labels.
For example, the first feature function in the first feature function set is tk(yi，yi-1X, i) where k is 1,2,3, …, and the second set of feature functions has s as the second feature functionl(yiX, i), wherein l ═ 1,2,3, …, wherein:
the conditional probability function P (y | x) is:
wherein λ isk，μlAre respectively a first characteristic function tk(yi，yi-1X, i) and a second characteristic function sl(yiWeight of x, i)The value may be a predetermined value.
The conditional probability function can be constructed according to equations (3) and (4). Then, the term sequence obtained by splitting the evaluation information may be used as the observation sequence x in the conditional probability function, and the probability of whether each term corresponds to an evaluator constituting the target evaluation object may be calculated, so as to obtain the probability sequence of the evaluator constituting the target evaluation object corresponding to the term sequence.
In this embodiment, each word may be subjected to corresponding weighting processing according to the probability that each word constitutes an evaluation target for the target evaluation object obtained in step 201, and then input to the trained emotion classification model for classification. The higher the probability that the word constitutes the evaluation word for the target evaluation object, the higher the corresponding weight. In this way, the weight of the word constituting the evaluation term for the target object in the input data of the trained emotion classification model is high, and the influence thereof on the emotion classification result is also large.
The emotion classification model can classify the emotional tendency of the input evaluation information. Here, the emotion categories may be predefined categories, for example, including a positive emotion category and a negative emotion category, or may be categories determined by a more detailed division manner, for example, a strong positive emotion category, a general positive emotion category, a slight negative emotion category, a strong negative emotion category, and the like. The embodiment of the disclosure does not specially limit the dividing mode of the emotion categories.
The emotion classification model can be obtained by training in advance based on the sample evaluation information labeled with the emotion category aiming at the target evaluation object.
Optionally, the evaluation information may be classified as follows: and converting words in the evaluation information into word vectors, multiplying the word vectors by the corresponding probability of forming the evaluation words aiming at the target evaluation object, inputting the trained emotion classification model for classification, and obtaining emotion category information aiming at the target evaluation object corresponding to the evaluation information.
In the implementation manner, the word sequence obtained by splitting the evaluation information may be vectorized by a word-to-vector method, converted into a word vector sequence, and multiplied by the probability corresponding to each word in the word sequence obtained in step 201 to serve as the input data of the emotion classification model, so as to increase the influence weight of words with a high probability of constituting the evaluation word for the target evaluation object in the emotion classification, and decrease the influence weight of words with a low probability of not constituting the evaluation word for the target evaluation object in the emotion classification.
Further optionally, the trained emotion classification model comprises a long-short term memory network and a classifier. The product of a word vector sequence obtained after the word sequence conversion and a corresponding probability sequence forming an evaluation word aiming at a target evaluation object can be used as the input of a long-short term memory network, the long-short term memory network is used for calculating a corresponding hidden state sequence for the input sequence, and then the hidden state is input into a classifier for classification, so that the input word sequence is mapped to a specific emotion category label. The emotion type label may be a label preset according to different emotion types, for example, a label of a positive emotion type may be preset to be 1, and a label of a negative emotion type may be preset to be 0.
The computation logic of the long-short term memory network is shown in fig. 3. Wherein, ctIs a hidden state, called the cellular state, it，ft，otAn input gate, a forgetting gate and an output gate. Input door itFor the input of control data, it is responsible for the input of the current position in the sequence (t denotes the current position, t is 1,2,3, …, n, n is the length of the input sequence), forgetting gate ftA gate O for controlling the forgetting of data, a gate O for controlling the state of the cell in the previous layer with a certain probabilitytFor controlling the output of data. The calculation formula is as follows:
it＝σ(Wixt+Uiht-1+bi)， (5)
ft＝σ(Wfxt+Ufht-1+bf)， (6)
ot＝σ(Woxt+Uoht-1+bo)， (7)
wherein, Wi，UiIs the weight parameter of the input gate, biIs the bias parameter of the input gate; wf，UfIs a weight parameter of the forgotten door, bfIs a biasing parameter of the forgetting gate; wo，UoIs the weight parameter of the output gate, boIs the offset parameter of the output gate; wc，UcIs a weight parameter of a hidden state, bcIs a bias parameter for the hidden state; σ is an activation function, which may be a sigmod function;
hadamard product (Hadamard product), x, representing a matrixtIs the element at the t-th position in the input sequence, htOutput for long-short term memory network corresponding to xtImplicit state of (2). Wherein, Wi，Ui，bi，Wf，Uf，bf，Wo，Uo，bo，Wc，Uc，bcAre parameters that require training.
The classifier can be constructed based on a softmax function (normalized exponential function), and the probability of the evaluation information for the emotion category of the target evaluation object can be determined according to the following formula (11):
wherein L isiIndicates the probability that the emotion type of the evaluation information for the target evaluation object is the i-th set emotion type, wiThe weighting parameter of the emotion category set for the ith, i ═ 1,2,3, …, K; k is the total number of emotion classes set, hnThe output implicit state corresponding to the last (nth) word of the input sequence in the long-short term memory network is b, the bias parameter in the softmax function is wiAnd b is a parameter that requires training.
From the calculation result of the above equation (11), the probability value L can be determined from the set emotion classificationiThe maximum emotion type is emotion type information for the target evaluation object corresponding to the evaluation information.
With continued reference to FIG. 4, a schematic diagram of an implementation of a method for determining sentiment categories of rating information according to the present application is shown.
As shown in fig. 4, after the evaluation information is divided into word sequences X1, X2, X3, …, Xn, corresponding conditional probabilities K1, K2, K3, …, Kn are calculated by using a conditional random field, words in the word sequences X1, X2, X3, …, Xn are vectorized and multiplied by corresponding conditional probabilities X1, X2, X3, …, Xn respectively, and then input LSTM (Long Short-Term Memory network), and then the hidden state of the last stage output gate is input into a classifier constructed based on a softmax function, so as to obtain an emotion tag for a target evaluation object.
The identification result aiming at the emotion types of the target evaluation objects can help to quickly classify the target evaluation objects, and in an actual scene, evaluation information can be gathered according to different emotion types and pushed to interested users. For example, the user can view good evaluation or poor evaluation aiming at logistics in a classified manner, and the user who uses the logistics service does not need to select tags such as 'poor evaluation' or 'good evaluation' when submitting evaluation information, so that intelligent evaluation information emotion classification is realized.
Optionally, the method for determining the emotion category of the evaluation information may further include: and training based on a sample evaluation information set to obtain a trained emotion classification model, wherein the sample evaluation information set comprises sample evaluation information and marking information of the sample evaluation information aiming at the emotion type of the target evaluation object.
The sample rating information may be data collected from the internet containing rating content. The corresponding marking information can be obtained by manually judging the emotion types and marking. During training, the sample evaluation information can be input into the emotion classification model to be trained to obtain the emotion classification result of the sample evaluation information, and the parameter of the emotion classification model to be trained is iteratively adjusted by comparing the emotion classification result of the sample evaluation information with the difference between the sample evaluation information and the label information of the sample evaluation information for the emotion category of the target evaluation object, so that the difference between the emotion classification result of the emotion classification model to be trained for the sample evaluation information and the label information of the sample evaluation information for the emotion category of the target evaluation object is gradually reduced to a certain range.
The process of the method for determining the emotion classification of the evaluation information according to the above embodiment of the present application calculates the probability that the words contained in the evaluation information constitute the evaluation words for the target evaluation object based on the preset evaluation keyword library of the target evaluation object, performs emotion classification on the evaluation information by using the trained emotion classification model based on the probability that the words contained in the evaluation information constitute the evaluation words for the target evaluation object, obtains emotion classification information corresponding to the evaluation information for the target evaluation object, realizes the introduction of an attention mechanism in emotion classification of the evaluation information, can classify the emotions of different evaluation objects for the evaluation information, the method can avoid the mutual influence of the emotional tendency judgment results aiming at different evaluation objects in the same evaluation information, and can obtain more accurate evaluation information emotion classification results.
In addition, in some implementation modes, the conditional random field is adopted to estimate the conditional probability that each word in the evaluation information forms the evaluation language aiming at the target evaluation object, and the feature function which accords with the evaluation language feature of the target evaluation object can be limited in the conditional random field, so that the emotion type recognition accuracy is improved.
In some implementation modes, the long-term and short-term memory network and the classifier are used as the emotion classification model, the emotion classification of the evaluation information aiming at the target evaluation object can be effectively determined by combining the context information, and the accuracy of the emotion classification recognition result is improved.
With continued reference to FIG. 5, a flow diagram of another embodiment of a method for determining sentiment categories for rating information in accordance with the present disclosure is shown. As shown in fig. 5, a flowchart 500 of the method for determining an emotion category of evaluation information according to the present embodiment includes the following steps:
And 502, performing emotion classification on the evaluation information by adopting a trained emotion classification model based on the probability that the words contained in the evaluation information form an evaluation word aiming at the target evaluation object to obtain emotion category information corresponding to the evaluation information aiming at the target evaluation object.
Step 501 and step 502 of this embodiment are respectively the same as step 201 and step 202 of the foregoing embodiment, and specific implementation manners of step 501 and step 502 may refer to descriptions of step 201 and step 202 of the foregoing embodiment, which are not described herein again.
The evaluation information which is preset emotion type information and is aimed at the emotion type information of the target evaluation object can be screened out as the evaluation information to be pushed according to the emotion type information of the target evaluation object of each evaluation information obtained in the step 502.
The content providing subject associated with the target evaluation object may be a content providing subject which takes the target evaluation object as service content, specifically, a subject which produces, manufactures or edits content detail information of the target evaluation object, and the preset emotion category information may be emotion category information subscribed by the content providing subject of the target evaluation object.
In an exemplary scenario, the target evaluation object may be a commodity of an online shopping platform, and the content provider of the target evaluation object may be a manufacturer of the target evaluation object, a store user who provides commodity purchase information and detail information. The preset emotion category information may be, for example, information characterizing negative emotion tendencies. By pushing the evaluation information to be pushed to the content providing main body interested in the preset emotion category information, the content providing main body associated with the target evaluation object can be provided with accurate evaluation feedback information, so that the content providing main body is helped to adjust the content (namely the target evaluation object) provided by the content providing main body based on the evaluation information, such as adjusting stock, changing the design of a product, and the like.
With further reference to fig. 6, as an implementation of the methods shown in the above-mentioned figures, the present application provides an embodiment of an apparatus for determining an emotion category of evaluation information, which corresponds to the method embodiments shown in fig. 2 and fig. 5, and which is particularly applicable to various electronic devices.
As shown in fig. 6, the apparatus 600 for determining emotion category of evaluation information of the present embodiment may include a calculation unit 601 and a classification unit 602. The calculation unit 601 is configured to calculate the probability that a word included in the evaluation information constitutes an evaluation word for the target evaluation object based on a preset evaluation keyword library of the target evaluation object; the classification unit 602 is configured to perform emotion classification on the evaluation information by using a trained emotion classification model based on the probability that the words included in the evaluation information constitute the evaluation words for the target evaluation object, and obtain emotion category information corresponding to the evaluation information and for the target evaluation object.
In some embodiments, the above-mentioned calculating unit 601 may be further configured to calculate the probability that the word included in the evaluation information constitutes the evaluation word for the target evaluation object as follows: splitting the evaluation information into word sequences; constructing a feature function set of a conditional random field based on a preset evaluation keyword library of a target evaluation object, and constructing a conditional probability function based on the feature function set of the conditional random field, wherein the conditional probability function represents the probability of outputting a corresponding evaluation word tag sequence when a text word sequence is input in the conditional random field, and an evaluation word tag in the evaluation word tag sequence is used for identifying that a corresponding input text word forms an evaluation word aiming at the target evaluation object or does not form the evaluation word aiming at the target evaluation object; and calculating the probability that each word in the word sequence forms the appraisal words aiming at the target appraisal object based on the conditional probability function.
In some embodiments, the feature function set includes a first feature function set and a second feature function set, where the first feature function in the first feature function set is used to characterize transfer features between adjacent evaluator labels corresponding to adjacent text words in the input text word sequence, and the second feature function in the second feature function set is used to characterize mapping features of words in the input text word sequence mapped to corresponding evaluator labels; the above-mentioned calculation unit 601 may be further configured to construct the conditional probability function as follows: a conditional probability function is constructed based on the weighted sum of the first feature functions and the weighted sum of the second feature functions.
In some embodiments, the classifying unit 602 may be further configured to perform emotion classification on the evaluation information to obtain emotion category information corresponding to the evaluation information for the target evaluation object as follows: and converting words in the evaluation information into word vectors, multiplying the word vectors by the corresponding probability of forming the evaluation words aiming at the target evaluation object, inputting the trained emotion classification model for classification, and obtaining emotion category information aiming at the target evaluation object corresponding to the evaluation information.
In some embodiments, the trained emotion classification model includes a long-short term memory network and a classifier; the apparatus 600 may further include: and the training unit is configured to train to obtain a trained emotion classification model based on a sample evaluation information set, wherein the sample evaluation information set comprises sample evaluation information and marking information of the sample evaluation information aiming at the emotion category of the target evaluation object.
In some embodiments, the apparatus 600 may further include: the pushing unit is configured to determine that the emotion category information of the target evaluation object is evaluation information of preset emotion category information, the evaluation information serves as evaluation information to be pushed, and the evaluation information to be pushed is pushed to a content providing main body related to the target evaluation object in the evaluation information to be pushed.
It should be understood that the elements described in apparatus 600 correspond to various steps in the methods described with reference to fig. 2 and 5. Thus, the operations and features described above for the method are equally applicable to the apparatus 600 and the units included therein, and are not described in detail here.
The apparatus for determining emotion classification of evaluation information according to the above embodiment of the present application calculates, by using a conditional random field, a probability that a word included in the evaluation information constitutes an evaluation phrase for a target evaluation object, performs emotion classification on the evaluation information by using a trained emotion classification model based on the probability that the word included in the evaluation information constitutes the evaluation phrase for the target evaluation object, obtains emotion classification information for the target evaluation object corresponding to the evaluation information, and can separate emotion classification for different evaluation objects when performing emotion classification on the evaluation information, thereby obtaining a more accurate evaluation information emotion classification result.
Referring now to FIG. 7, a block diagram of an electronic device (e.g., the server of FIG. 1) 700 suitable for use in implementing embodiments of the present disclosure is shown. The server shown in fig. 7 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.
As shown in fig. 7, electronic device 700 may include a processing means (e.g., central processing unit, graphics processor, etc.) 701 that may perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)702 or a program loaded from storage 708 into a Random Access Memory (RAM) 703. In the RAM 703, various programs and data necessary for the operation of the electronic apparatus 700 are also stored. The processing device 701, the ROM 702, and the RAM 703 are connected to each other by a bus 704. An input/output (I/O) interface 705 is also connected to bus 704.
The following devices may be connected to the I/O interface 705: input devices 706 including, for example, a touch screen, touch pad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; an output device 707 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; a storage device 708 including, for example, a hard disk; and a communication device 709. The communication means 709 may allow the electronic device 700 to communicate wirelessly or by wire with other devices to exchange data. While fig. 7 illustrates an electronic device 700 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided. Each block shown in fig. 7 may represent one device or may represent multiple devices as desired.
In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such embodiments, the computer program may be downloaded and installed from a network via the communication means 709, or may be installed from the storage means 708, or may be installed from the ROM 702. The computer program, when executed by the processing device 701, performs the above-described functions defined in the methods of embodiments of the present disclosure. It should be noted that the computer readable medium described in the embodiments of the present disclosure may be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In embodiments of the disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In embodiments of the present disclosure, however, a computer readable signal medium may comprise a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.
The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device. The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: calculating the probability that the words contained in the evaluation information form the evaluation words aiming at the target evaluation object based on a preset evaluation keyword library of the target evaluation object; and performing emotion classification on the evaluation information by adopting a trained emotion classification model based on the probability that the words contained in the evaluation information form an evaluation word aiming at the target evaluation object to obtain emotion category information corresponding to the evaluation information and aiming at the target evaluation object.
Computer program code for carrying out operations for embodiments of the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The units described in the embodiments of the present application may be implemented by software or hardware. The described units may also be provided in a processor, and may be described as: a processor includes a calculation unit and a classification unit. Here, the names of these units do not constitute a limitation to the unit itself in some cases, and for example, the calculation unit may also be described as "a unit that calculates the probability that a word included in the evaluation information constitutes an evaluator for the target evaluation object based on a preset evaluation keyword library of the target evaluation object".
The above description is only a preferred embodiment of the application and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the invention herein disclosed is not limited to the particular combination of features described above, but also encompasses other arrangements formed by any combination of the above features or their equivalents without departing from the spirit of the invention. For example, the above features may be replaced with (but not limited to) features having similar functions disclosed in the present application.
Claims (9)
1. A method for determining an emotion classification for rating information, comprising:
calculating the probability that the words contained in the evaluation information form the evaluation words aiming at the target evaluation object based on a preset evaluation keyword library of the target evaluation object;
and carrying out emotion classification on the evaluation information by adopting a trained emotion classification model based on the probability that the words contained in the evaluation information form an evaluation word aiming at the target evaluation object, so as to obtain emotion category information corresponding to the evaluation information and aiming at the target evaluation object.
2. The method according to claim 1, wherein the calculating the probability that the words included in the evaluation information constitute the evaluation words for the target evaluation object based on the preset evaluation keyword library of the target evaluation object comprises:
splitting the evaluation information into word sequences;
constructing a feature function set of a conditional random field based on a preset evaluation keyword library of a target evaluation object, and constructing a conditional probability function based on the feature function set of the conditional random field, wherein the conditional probability function represents the probability of outputting a corresponding evaluation word tag sequence when a text word sequence is input in the conditional random field, and the evaluation word tag in the evaluation word tag sequence is used for identifying that the corresponding input text word forms an evaluation language for the target evaluation object or does not form the evaluation language for the target evaluation object;
and calculating the probability that each word in the word sequence forms an appraisal word aiming at a target appraisal object based on the conditional probability function.
3. The method of claim 2, wherein the set of feature functions comprises a first set of feature functions and a second set of feature functions, wherein a first feature function in the first set of feature functions is used for characterizing transfer features between adjacent rater labels corresponding to adjacent text words in the input text word sequence, and a second feature function in the second set of feature functions is used for characterizing mapping features of words in the input text word sequence mapped to corresponding rater labels;
the method for constructing the conditional probability function based on the feature function set of the conditional random field comprises the following steps:
and constructing the conditional probability function based on the weighted sum of the first characteristic functions and the weighted sum of the second characteristic functions.
4. The method according to claim 1, wherein the obtaining of emotion classification information for a target evaluation object corresponding to evaluation information by emotion classification of the evaluation information using a trained emotion classification model based on a probability that a word included in the evaluation information constitutes an evaluation word for the target evaluation object comprises:
and converting the words in the evaluation information into word vectors, multiplying the word vectors by the corresponding probability of forming the evaluation words aiming at the target evaluation object, inputting the trained emotion classification model for classification, and obtaining emotion category information aiming at the target evaluation object corresponding to the evaluation information.
5. The method of claim 4, wherein the trained emotion classification model comprises a long-short term memory network and a classifier;
the method further comprises the following steps:
and training based on a sample evaluation information set to obtain the trained emotion classification model, wherein the sample evaluation information set comprises sample evaluation information and marking information of the sample evaluation information aiming at the emotion type of a target evaluation object.
6. The method of any of claims 1-5, wherein the method further comprises:
determining that the emotion category information of the target evaluation object is evaluation information of preset emotion category information, using the evaluation information as evaluation information to be pushed, and pushing the evaluation information to be pushed to a content providing main body associated with the target evaluation object in the evaluation information to be pushed.
7. An apparatus for determining an emotion classification for rating information, comprising:
a calculation unit configured to calculate a probability that a word included in the evaluation information constitutes an evaluation word for the target evaluation object based on a preset evaluation keyword library of the target evaluation object;
and the classification unit is configured to perform emotion classification on the evaluation information by adopting a trained emotion classification model based on the probability that the words contained in the evaluation information form an evaluation word aiming at a target evaluation object, so as to obtain emotion category information corresponding to the evaluation information and aiming at the target evaluation object.
8. An electronic device, comprising:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-6.
9. A computer-readable medium, on which a computer program is stored, wherein the program, when executed by a processor, implements the method of any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910417377.XA CN111966822A (en) | 2019-05-20 | 2019-05-20 | Method and device for determining emotion category of evaluation information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910417377.XA CN111966822A (en) | 2019-05-20 | 2019-05-20 | Method and device for determining emotion category of evaluation information |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111966822A true CN111966822A (en) | 2020-11-20 |
Family
ID=73357823
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910417377.XA Pending CN111966822A (en) | 2019-05-20 | 2019-05-20 | Method and device for determining emotion category of evaluation information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111966822A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102682124A (en) * | 2012-05-16 | 2012-09-19 | 苏州大学 | Emotion classifying method and device for text |
| CN104298665A (en) * | 2014-10-16 | 2015-01-21 | 苏州大学 | Identification method and device of evaluation objects of Chinese texts |
| CN106156004A (en) * | 2016-07-04 | 2016-11-23 | 中国传媒大学 | The sentiment analysis system and method for film comment information based on term vector |
| CN107102980A (en) * | 2016-02-19 | 2017-08-29 | 北京国双科技有限公司 | The extracting method and device of emotion information |
| CN107544957A (en) * | 2017-07-05 | 2018-01-05 | 华北电力大学 | A kind of Sentiment orientation analysis method of business product target word |
| CN109086340A (en) * | 2018-07-10 | 2018-12-25 | 太原理工大学 | Evaluation object recognition methods based on semantic feature |
| CN109241993A (en) * | 2018-07-26 | 2019-01-18 | 中国科学院自动化研究所 | Merge the evaluation object sensibility classification method and device of user and overall evaluation information |
-
2019
- 2019-05-20 CN CN201910417377.XA patent/CN111966822A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102682124A (en) * | 2012-05-16 | 2012-09-19 | 苏州大学 | Emotion classifying method and device for text |
| CN104298665A (en) * | 2014-10-16 | 2015-01-21 | 苏州大学 | Identification method and device of evaluation objects of Chinese texts |
| CN107102980A (en) * | 2016-02-19 | 2017-08-29 | 北京国双科技有限公司 | The extracting method and device of emotion information |
| CN106156004A (en) * | 2016-07-04 | 2016-11-23 | 中国传媒大学 | The sentiment analysis system and method for film comment information based on term vector |
| CN107544957A (en) * | 2017-07-05 | 2018-01-05 | 华北电力大学 | A kind of Sentiment orientation analysis method of business product target word |
| CN109086340A (en) * | 2018-07-10 | 2018-12-25 | 太原理工大学 | Evaluation object recognition methods based on semantic feature |
| CN109241993A (en) * | 2018-07-26 | 2019-01-18 | 中国科学院自动化研究所 | Merge the evaluation object sensibility classification method and device of user and overall evaluation information |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Duchi et al. | Distributionally robust losses for latent covariate mixtures | |
| CN107679234B (en) | Customer service information providing method, customer service information providing device, electronic equipment and storage medium | |
| US20180197087A1 (en) | Systems and methods for retraining a classification model | |
| EP3547155A1 (en) | Entity representation learning for improving digital content recommendations | |
| CN112115257A (en) | Method and apparatus for generating information evaluation model | |
| CN110555469A (en) | Method and device for processing interactive sequence data | |
| CN115917535A (en) | Recommendation model training method, recommendation device and computer readable medium | |
| CN111783039A (en) | Risk determination method, risk determination device, computer system and storage medium | |
| CN111783810A (en) | Method and apparatus for determining attribute information of user | |
| CN111667024B (en) | Content pushing method, device, computer equipment and storage medium | |
| Gil-Gonzalez et al. | Learning from multiple annotators using kernel alignment | |
| US20190378043A1 (en) | Technologies for discovering specific data in large data platforms and systems | |
| CN117217284A (en) | Data processing method and device | |
| Ertekin et al. | Approximating the crowd | |
| CN111191677A (en) | User characteristic data generation method and device and electronic equipment | |
| Li et al. | The user preference identification for product improvement based on online comment patch | |
| CN110826327A (en) | Emotion analysis method and device, computer readable medium and electronic equipment | |
| CN113032676B (en) | Recommendation method and system based on micro-feedback | |
| Latha et al. | Product recommendation using enhanced convolutional neural network for e-commerce platform | |
| He et al. | Semi-supervised dialogue abstractive summarization via high-quality pseudolabel selection | |
| CN113496236B (en) | User tag information determining method, device, equipment and storage medium | |
| JP2024522358A (en) | Machine Learning Assisted Automated Taxonomies for Marketing Automation and Customer Relationship Management Systems | |
| CN111966822A (en) | Method and device for determining emotion category of evaluation information | |
| CN112200602B (en) | Neural network model training method and device for advertisement recommendation | |
| CN113792952A (en) | Method and apparatus for generating a model |
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 |