CN113761910A - Comment text fine-grained emotion analysis method integrating emotional characteristics - Google Patents
Comment text fine-grained emotion analysis method integrating emotional characteristics Download PDFInfo
- Publication number
- CN113761910A CN113761910A CN202110283681.7A CN202110283681A CN113761910A CN 113761910 A CN113761910 A CN 113761910A CN 202110283681 A CN202110283681 A CN 202110283681A CN 113761910 A CN113761910 A CN 113761910A
- Authority
- CN
- China
- Prior art keywords
- emotion
- fine
- vector
- grained
- comment
- 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
- 230000008451 emotion Effects 0.000 title claims abstract description 65
- 238000004458 analytical method Methods 0.000 title claims abstract description 24
- 230000002996 emotional effect Effects 0.000 title claims description 10
- 239000013598 vector Substances 0.000 claims abstract description 56
- 238000012549 training Methods 0.000 claims abstract description 21
- 238000007781 pre-processing Methods 0.000 claims abstract description 10
- 238000000034 method Methods 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 2
- 238000013527 convolutional neural network Methods 0.000 abstract description 18
- 230000007246 mechanism Effects 0.000 abstract description 9
- 238000011156 evaluation Methods 0.000 abstract description 8
- 230000002787 reinforcement Effects 0.000 abstract description 5
- 230000006870 function Effects 0.000 description 8
- 230000008569 process Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 235000012054 meals Nutrition 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- 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
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/211—Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Abstract
The invention discloses a comment text fine-grained emotion analysis method fused with emotion characteristics, which comprises the following steps of S1, preprocessing the existing comment data corpus; s2, constructing a joint vector; and S3, carrying out ADBC fine-grained emotion model training. Has the advantages that: the invention represents the text vector by fusing the emotion label, the emotion word and the aspect word, designs a high-robustness fine-grained emotion analysis framework aiming at the enhancement semantics of the comment text, more excavates the potential emotion information in the text, and simultaneously embeds the CNN convolutional neural network to carry out feature reinforcement learning on the output layer of the Bi-GRU before the attention mechanism, thereby improving the accuracy of fine-grained emotion analysis. The real evaluation information of the user can be reflected, so that the user can more directly know the quality of each aspect of the product. The feedback result is provided for the consumers and the suppliers with high efficiency and reliability.
Description
Technical Field
The invention relates to the technical field of natural language text processing, in particular to a comment text fine-grained emotion analysis method integrating emotion characteristics.
Background
With the development and perfection of the network and the APP terminal, people can conveniently conduct online shopping, sharing and commenting on concerned things, for example, after people buy commodities, people habitually comment on various attributes of the commodities, describe the actual conditions of the commodities and describe the use experience of the people. After a meal is received at a restaurant, dishes, services, etc. are reviewed. Through the emotion analysis on the comments, the emotional tendency and the satisfaction degree of the user on the commodity can be obtained, and the production, sale and purchase decisions of people can be helped.
The fine-grained sentiment analysis focuses on the mining of attribute level information, and generally requires that comment texts have strong sentiment tendency or definite attribute information, so that the information extraction of related texts is effectively enhanced by constructing a domain sentiment dictionary.
In the previous fine-grained sentiment analysis, most of the sentiment tendency of the attribute is determined by extracting attribute and viewpoint pairs, namely the attribute and the viewpoint are determined by the appearance of pairs, and the sentiment determination accuracy of the attribute level is low; this also tends to create a lack of views of some implicit attributes, such as: (1) the catering method comprises the following steps: "relatively close, very convenient", there is no attribute, but is an evaluation to "position"; (2) and (3) evaluation aspects: the expression of 'very noisy sound' is the evaluation of 'sound effect'; (3) and (3) on-line purchase: "with damage" expresses an assessment of "integrity of the good". On the other hand, there is a possibility that the evaluation of the product or the attribute is not performed even if there is an evaluation viewpoint, and therefore, there is a lack of a method for improving the accuracy of the judgment of the fine-grained emotional tendency by using a comment sentence having emotion.
An effective solution to the problems in the related art has not been proposed yet.
Disclosure of Invention
The invention provides a fine-grained sentiment analysis framework facing product comments, and provides a fine-grained sentiment analysis method based on Att-Dic-BiGRU-CNN (ADBC). by means of the idea of mapping label information to semantic space, in the fine-grained sentiment analysis, sentiment aspect word information, sentiment word information, corresponding part-of-speech information and word vectors are used as an embedding layer, a Bi-GRU network is used for carrying out sentiment semantic information feature extraction on information in comment text context, then the information is input to a convolutional neural network, the information is output to a sentiment mechanism layer introducing sentiment labels for weight distribution, and finally label classification is carried out on text features through a softmax function. The method can effectively improve the accuracy of fine-grained emotion classification of products.
In order to achieve the purpose, the invention provides the following technical scheme:
a comment text fine-grained emotion analysis method integrating emotion characteristics comprises the following steps:
s1, preprocessing the linguistic data of the existing comment data;
s2, constructing a joint vector;
and S3, carrying out ADBC fine-grained emotion model training.
Further, the pre-processing of the comment data corpus in the step S1 includes the following steps:
s11, pre-processing the comment corpus, such as word segmentation, stop word removal, punctuation and the like;
s12, using random initialization vector matrix for characters, attribute terms, emotional terms and the like in the sentence;
s13, constructing an aspect word vector sequence for the attribute aspect words;
s14, constructing an emotion word vector sequence for the emotion words;
and S15, constructing a label vector for the emotion labels.
Further, the step S2 is to construct a joint vector, and combine the plurality of vectors in the step S1, and adopt a concat function in the tensoflow to splice the vector sequences, so as to obtain a new fusion vector model.
Further, the step S3ADBC fine-grained emotion model training includes the following steps:
s31, training a GRU network layer;
s32, training a CNN network layer;
s33, and training an attention network layer.
Compared with the prior art, the invention has the following beneficial effects: according to the invention, by fusing external emotion characteristics, more potential emotion information in the text is mined, and meanwhile, a CNN convolutional neural network is embedded before an attention mechanism to perform characteristic reinforcement learning on the output layer of the Bi-GRU, so that the accuracy of fine-grained emotion analysis is improved. The invention represents the text vector by fusing the emotion label, the emotion word and the aspect word, designs a high-robustness fine-grained emotion analysis framework aiming at the enhancement semantics of the comment text, more excavates the potential emotion information in the text, and simultaneously embeds the CNN convolutional neural network to carry out feature reinforcement learning on the output layer of the Bi-GRU before the attention mechanism, thereby improving the accuracy of fine-grained emotion analysis. The real evaluation information of the user can be reflected, so that the user can more directly understand the product in all aspects. The feedback result is provided for the consumers and the suppliers with high efficiency and reliability.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flowchart of a comment text fine-grained emotion analysis method fused with emotion characteristics according to an embodiment of the present invention;
FIG. 2 is a vector diagram of the comment text fine-grained emotion analysis method fused with emotion features according to the embodiment of the invention.
Detailed Description
The invention is further described with reference to the following drawings and detailed description:
referring to fig. 1, a comment text fine-grained emotion analysis method fused with emotion features according to an embodiment of the present invention includes the following steps:
s1, preprocessing the linguistic data of the existing comment data;
s2, constructing a joint vector;
and S3, carrying out ADBC fine-grained emotion model training.
Preprocessing the corpus of the comment data in the step S1;
and S11, performing preprocessing such as word segmentation, stop word removal, punctuation and the like on the comment corpus.
S12, performing word vector representation on the comment text by using a random initialization vector matrix for characters, attribute terms, emotional words and the like in the sentence:
xc＝[c1,c2,...,cn]
s13, constructing an aspect word vector sequence for the attribute aspect words, wherein the sequence is expressed as:
xw＝[w1,w2,...,wm]
s14, constructing an emotion word vector sequence for the emotion words, wherein the emotion word vector sequence is expressed as:
xs＝[s1,s2,...,sm]
s15, constructing a label vector for the emotion labels, wherein the label vector is expressed as:
xl＝[l1,l2,...,lm]
the step S2 joint vector construction;
combining the multiple vectors in step 1, splicing the vector sequences by using a concat function in tenserflow, specifically referring to the process of fig. 2, to obtain a new fusion vector model, which can be expressed as:
xt＝concat([xw,xs,xc],1)
s3, carrying out ADBC fine-grained emotion model training;
s31, training a GRU network layer;
and (4) taking the joint vector obtained in the step (S2) as an input of the BiGRU network, namely performing vector language representation according to the sequence of words in the sentence, and coding the sentence. The method is characterized in that a weight is learned for a vector of each time step of a hidden layer, namely different weights are given to different words in a comment text when the vector representation of a sentence is obtained, and then the vector representation of the sentence is obtained by weighting the combined vectors of the different weights. The input and output layers represent the process shown in fig. 2. Model training was performed using the ADBC model, where the formula for the GRU is expressed as follows:
forget gate representation: z is a radical oft＝σ(Wz·[ht-1,xt]+bz)
The output gate represents: r ist＝σ(Wr·[ht-1,xt]+br)
Hidden layer representation:
the memory cell represents:
s32, training a CNN network layer;
convolutional neural networks can learn to capture a particular feature regardless of its location. Thus, the hidden unit is then input as a CNN convolutional layer, the sentence being:
H＝h0+h1+...+hn-1
vector hi:i+jRepresenting a series of word vectors hi,hi+1,...,hi+j. Each convolution operation has a filter w that generates a new signature through a window containing m words:
Hi＝f(w*hi:i+m-1+b)
where b is the bias, the final text vector is represented as:
H*＝[H0,H1,...,Hn-m]
s33, and training an attention network layer.
Adding an Attention mechanism for embedding emotion tags on the basis of a CNN network, and calculating cosine similarity between tag vector and each word in the characteristics of a sequence sentence output by a CNN network, namely the product of L2norm of the c-th tag embedding and the L-th word embedding. Wherein is a normalized matrix:
to better capture information between successive words, non-linearities are introduced in the similarity, here implemented jointly using convolution and activation functions, denoted gl-r:l+rThe relevance of the central word l in a window with the length of 2r +1 obtains the similarity between the central word l and each label:
ul＝RELU(gl-r:l+rW1+b1)
through maximum pooling, the label with the maximum correlation with the word l and the corresponding similarity can be obtained:
ml＝f(ul)
after obtaining the maximum similarity between each word in the article and the label, m is a vector with the length of L, and after softmax, the attention score of the normalized text sequence of the word is obtained as follows:
β＝softmax(ml)
and carrying out weighted sum on word vectors of each word in the article, wherein the text represents sentence emotional tendency in fine granularity obtained by label-based attribute score weighted word embedding:
the attention mechanism is represented by an additive model:
a＝softmax(w(t)c(t))
the final sentence vector is represented as:
v＝tanh(h*a)
then judging the emotional tendency, and taking the output v of the Attention mechanism as the input of the softmax layer to perform fine-grained emotional classification prediction vector:
y＝softmax(w*v+b)
and obtaining a final judgment result vector of the sentence. The vector length is the number of types of the judgment sentences, the vector values correspond to the positive, neutral and negative emotion numerical values respectively, and the emotion type corresponding to the maximum value is the emotion tendency of the fine granularity level.
The loss value is calculated using a cross entropy loss function:
E＝-∑y+λ||θ||2
using random gradient descent method for the model parameters. In the training, iterative training is carried out through the set minimum batch, the corpus in each minipatch calculates a loss function, and parameters in the network are optimized through back propagation. In the gradient descent process, dropout and an L2 paradigm are adopted to constrain the weight vector. After multiple iterations, when the accuracy rate tends to be stable, the model training is completed. The loss function of the model puts the vector representation of the label into a classifier in addition to the cross entropy mentioned above, and adds the cross entropy of the probability distribution of the forward propagation result and the label to the loss function to obtain the vector representation of the supervised information training label.
Through the scheme, the potential emotion information in the text is more mined by fusing the external emotion characteristics, and the CNN convolutional neural network is embedded to perform feature reinforcement learning on the output layer of the Bi-GRU before the attention mechanism, so that the accuracy of fine-grained emotion analysis is improved.
In practical application, the emotion labels, the emotion words and the aspect words are fused to represent text vectors, a high-robustness fine-grained emotion analysis frame aiming at the enhancement semantics of comment texts is designed, more potential emotion information in the texts is mined, and meanwhile, a CNN (convolutional neural network) is embedded before the attention mechanism to conduct feature reinforcement learning on the output layer of the Bi-GRU, so that the accuracy of fine-grained emotion analysis is improved. The real evaluation information of the user can be reflected, so that the user can more directly know the quality of each aspect of the product. And providing efficient and reliable feedback results for consumers and suppliers.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (4)
1. A comment text fine-grained emotion analysis method fused with emotion characteristics is characterized by comprising the following steps:
s1, preprocessing the linguistic data of the existing comment data;
s2, constructing a joint vector;
and S3, carrying out ADBC fine-grained emotion model training.
2. The method for analyzing the comment text fine granularity sentiment fused with the sentiment features according to claim 1, wherein the pre-processing of the existing comment data corpus in the step S1 comprises the following steps:
s11, pre-processing the comment corpus, such as word segmentation, stop word removal, punctuation and the like;
s12, using random initialization vector matrix for characters, attribute terms, emotional terms and the like in the sentence;
s13, constructing an aspect word vector sequence for the attribute aspect words;
s14, constructing an emotion word vector sequence for the emotion words;
and S15, constructing a label vector for the emotion labels.
3. The method for fine-grained emotion analysis of comment text fused with emotion characteristics as claimed in claim 1, wherein said step S2 is a joint vector construction, and said vector sequence is spliced by using a concat function in tensoflow in combination with the plurality of vectors in step S1 to obtain a new fused vector model.
4. The method for analyzing the comment text fine-grained emotion fused with emotion characteristics according to claim 1, wherein the step S3 of ADBC fine-grained emotion model training comprises the following steps:
s31, training a GRU network layer;
s32, training a CNN network layer;
s33, and training an attention network layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110283681.7A CN113761910A (en) | 2021-03-17 | 2021-03-17 | Comment text fine-grained emotion analysis method integrating emotional characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110283681.7A CN113761910A (en) | 2021-03-17 | 2021-03-17 | Comment text fine-grained emotion analysis method integrating emotional characteristics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113761910A true CN113761910A (en) | 2021-12-07 |
Family
ID=78786733
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110283681.7A Pending CN113761910A (en) | 2021-03-17 | 2021-03-17 | Comment text fine-grained emotion analysis method integrating emotional characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113761910A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114756681A (en) * | 2022-04-28 | 2022-07-15 | 西安交通大学 | Evaluation text fine-grained suggestion mining method based on multi-attention fusion |
| CN116361472A (en) * | 2023-05-02 | 2023-06-30 | 周维 | Public opinion big data analysis system for social network comment hot events |
| CN116361472B (en) * | 2023-05-02 | 2024-05-03 | 脉讯在线(北京)信息技术有限公司 | Method for analyzing public opinion big data of social network comment hot event |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110083833A (en) * | 2019-04-18 | 2019-08-02 | 东华大学 | Term vector joint insertion sentiment analysis method in terms of Chinese words vector sum |
| CN110489523A (en) * | 2019-07-31 | 2019-11-22 | 西安理工大学 | A kind of fine granularity sentiment analysis method based on online shopping evaluation |
| CN110807320A (en) * | 2019-11-11 | 2020-02-18 | 北京工商大学 | Short text emotion analysis method based on CNN bidirectional GRU attention mechanism |
| CN111858945A (en) * | 2020-08-05 | 2020-10-30 | 上海哈蜂信息科技有限公司 | Deep learning-based comment text aspect level emotion classification method and system |
| CN112347258A (en) * | 2020-11-16 | 2021-02-09 | 合肥工业大学 | Short text aspect level emotion classification method |
-
2021
- 2021-03-17 CN CN202110283681.7A patent/CN113761910A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110083833A (en) * | 2019-04-18 | 2019-08-02 | 东华大学 | Term vector joint insertion sentiment analysis method in terms of Chinese words vector sum |
| CN110489523A (en) * | 2019-07-31 | 2019-11-22 | 西安理工大学 | A kind of fine granularity sentiment analysis method based on online shopping evaluation |
| CN110807320A (en) * | 2019-11-11 | 2020-02-18 | 北京工商大学 | Short text emotion analysis method based on CNN bidirectional GRU attention mechanism |
| CN111858945A (en) * | 2020-08-05 | 2020-10-30 | 上海哈蜂信息科技有限公司 | Deep learning-based comment text aspect level emotion classification method and system |
| CN112347258A (en) * | 2020-11-16 | 2021-02-09 | 合肥工业大学 | Short text aspect level emotion classification method |
Non-Patent Citations (1)
| Title |
|---|
| 季威志 等: "基于BiGRU-CNN-Attention模型的股市评论情感分析", 《工业控制计算机》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114756681A (en) * | 2022-04-28 | 2022-07-15 | 西安交通大学 | Evaluation text fine-grained suggestion mining method based on multi-attention fusion |
| CN114756681B (en) * | 2022-04-28 | 2024-04-02 | 西安交通大学 | Evaluation and education text fine granularity suggestion mining method based on multi-attention fusion |
| CN116361472A (en) * | 2023-05-02 | 2023-06-30 | 周维 | Public opinion big data analysis system for social network comment hot events |
| CN116361472B (en) * | 2023-05-02 | 2024-05-03 | 脉讯在线(北京)信息技术有限公司 | Method for analyzing public opinion big data of social network comment hot event |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108363695B (en) | User comment attribute extraction method based on bidirectional dependency syntax tree representation | |
| CN110245229B (en) | Deep learning theme emotion classification method based on data enhancement | |
| Li et al. | A deep learning-based approach to constructing a domain sentiment lexicon: a case study in financial distress prediction | |
| CN108446271B (en) | Text emotion analysis method of convolutional neural network based on Chinese character component characteristics | |
| CN110502626B (en) | Aspect level emotion analysis method based on convolutional neural network | |
| CN110765769B (en) | Clause feature-based entity attribute dependency emotion analysis method | |
| CN111339260A (en) | BERT and QA thought-based fine-grained emotion analysis method | |
| CN113392209B (en) | Text clustering method based on artificial intelligence, related equipment and storage medium | |
| CN112069320B (en) | Span-based fine-grained sentiment analysis method | |
| CN115952291B (en) | Financial public opinion classification method and system based on multi-head self-attention and LSTM | |
| US20220300735A1 (en) | Document distinguishing based on page sequence learning | |
| CN111738807B (en) | Method, computing device, and computer storage medium for recommending target objects | |
| CN113420212A (en) | Deep feature learning-based recommendation method, device, equipment and storage medium | |
| CN114648031A (en) | Text aspect level emotion recognition method based on bidirectional LSTM and multi-head attention mechanism | |
| CN112434514A (en) | Multi-granularity multi-channel neural network based semantic matching method and device and computer equipment | |
| CN114942974A (en) | E-commerce platform commodity user evaluation emotional tendency classification method | |
| Xu et al. | Research on the construction of crossborder e-commerce logistics service system based on machine learning algorithms | |
| CN113761910A (en) | Comment text fine-grained emotion analysis method integrating emotional characteristics | |
| CN115269768A (en) | Element text processing method and device, electronic equipment and storage medium | |
| CN107729509B (en) | Discourse similarity determination method based on recessive high-dimensional distributed feature representation | |
| CN115659990A (en) | Tobacco emotion analysis method, device and medium | |
| CN116029793A (en) | Commodity recommendation method, device, equipment and medium thereof | |
| CN115391570A (en) | Method and device for constructing emotion knowledge graph based on aspects | |
| CN115906824A (en) | Text fine-grained emotion analysis method, system, medium and computing equipment | |
| CN111666410B (en) | Emotion classification method and system for commodity user comment text |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211207 |
|
| RJ01 | Rejection of invention patent application after publication |