CN111695346B - Method for improving public opinion entity recognition rate in financial risk prevention and control field - Google Patents
Method for improving public opinion entity recognition rate in financial risk prevention and control field Download PDFInfo
- Publication number
- CN111695346B CN111695346B CN202010550784.0A CN202010550784A CN111695346B CN 111695346 B CN111695346 B CN 111695346B CN 202010550784 A CN202010550784 A CN 202010550784A CN 111695346 B CN111695346 B CN 111695346B
- Authority
- CN
- China
- Prior art keywords
- model
- entity
- training
- financial
- general
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 230000002265 prevention Effects 0.000 title claims abstract description 13
- 238000012549 training Methods 0.000 claims abstract description 60
- 238000000605 extraction Methods 0.000 claims abstract description 48
- 238000002372 labelling Methods 0.000 claims abstract description 28
- 238000013135 deep learning Methods 0.000 claims abstract description 18
- 238000003058 natural language processing Methods 0.000 claims description 17
- 238000013528 artificial neural network Methods 0.000 claims description 16
- 230000007246 mechanism Effects 0.000 claims description 12
- 238000012512 characterization method Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 4
- 238000013527 convolutional neural network Methods 0.000 claims description 4
- 125000004122 cyclic group Chemical group 0.000 claims description 4
- 230000001360 synchronised effect Effects 0.000 claims 1
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000012544 monitoring process Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 230000006872 improvement Effects 0.000 description 3
- 230000008520 organization Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
-
- 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
-
- 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
- G06F16/955—Retrieval from the web using information identifiers, e.g. uniform resource locators [URL]
- G06F16/9562—Bookmark management
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Machine Translation (AREA)
Abstract
The invention relates to the technical field of Internet, in particular to a method for improving the public opinion entity recognition rate in the financial risk prevention and control field, which comprises the following steps: s1, collecting general domain corpus, and labeling a general entity in the corpus with BIO labels; s2, collecting corpora in the financial field, and only labeling the financial abbreviations and the special entities of new financial entities in the corpora, and not labeling the general entities; s3, selecting a deep learning NLP pre-training model; s4, selecting a downstream model suitable for the NER task, S5, constructing an overall model for training the NER task, and training the overall model by using the general corpus collected in S1 to obtain a general entity extraction model. The method has the beneficial effects that the method based on deep learning is utilized, and the extraction results of the general field NER model based on the NLP pre-training model and the special field NER model based on the NLP pre-training model are respectively trained and then fused, so that the extraction rate of public opinion entity identification in the financial field is improved.
Description
Technical Field
The invention relates to the technical field of Internet, in particular to a method for improving the public opinion entity recognition rate in the financial risk prevention and control field.
Background
In the field of financial risk prevention and control, effective monitoring of network public opinion is required. Firstly, enterprises and products which are listed with a monitoring list need to be monitored, and secondly, new financial entities need to be discovered in time; in general, a named entity recognition technology in a natural language processing technology is used to extract entities belonging to an "organization" tag in public opinion as entities of the public opinion.
Chinese patent number 201610037682.2 provides a method and a device for analyzing public opinion event entity, which relate to the technical field of Internet and aim to solve the problems that a public opinion monitoring system cannot accurately analyze characters and mechanisms related to the public opinion event, so that a user cannot accurately position the source generated by the public opinion event through the public opinion monitoring system, and the optimal guiding mode for solving the public opinion event cannot be timely determined. The technical scheme of the invention comprises the following steps: acquiring an information set, and segmenting the information set; extracting character entities and mechanism entities in the segmented information set; counting the common mention times, the person entity mention times and the mechanism entity mention times respectively; determining the association relationship between the character entity and the mechanism entity according to the common reference times; and determining public opinion event entities and entity relationships according to the number of mention of the personage entities and/or the number of mention of the institution entities and the association relationship between the personage entities and the institution entities. The method and the device are applied to the process of monitoring the public opinion event.
However, because the general NER extraction model is formed based on general predictive training similar to encyclopedia and news, the model has strong extraction capability for common organization names, but has high difficulty for extracting short or new financial entities in the financial field, cannot be accurately identified, and causes information loss.
Disclosure of Invention
The invention aims to provide a method for improving the public opinion entity recognition rate in the financial risk prevention and control field, which aims to solve the problems that the extraction of short or new financial entities in the financial field has higher difficulty, cannot be accurately recognized and causes information loss in the background art.
The technical scheme of the invention is as follows: a method for improving the public opinion entity identification rate in the financial risk prevention and control field comprises the following steps:
s1, collecting general domain corpus, and labeling a general entity in the corpus with BIO labels;
S2, collecting corpora in the financial field, and only labeling the financial abbreviations and the special entities of new financial entities in the corpora, and not labeling the general entities;
S3, selecting a deep learning NLP pre-training model;
S4, selecting a downstream model suitable for the NER task;
S5, constructing an integral model for training NER tasks, training the integral model by using the general corpus collected in S1 to obtain a general entity extraction model, and training the integral model by using the financial domain-specific corpus collected in S2 to obtain a financial domain-specific entity extraction model;
S6, for public opinion needing to be extracted, respectively utilizing the two models obtained through training in the S5 to independently extract, wherein the general entity extraction model is responsible for extracting general entity, the new entity extraction model in the financial field is responsible for extracting entity short and new entity professional entity, and entity extraction results obtained by fusing the two models are obtained, so that the entity extraction results of the public opinion are obtained.
The method comprises the steps that in S3, NLP (natural language processing) is processed, understood and applied to human language by a computer, a pre-training model in S3 is trained through a large number of unlabeled language texts to obtain a set of model parameters, the model and corresponding parameters are pre-training models, in S3, deep learning is an algorithm for carrying out characterization learning on data by taking a multi-layer artificial neural network as a framework, a deep neural network, a convolutional neural network, a cyclic neural network and the like are common deep learning frameworks, in S4 NER is named entity identification, namely an entity with specific meaning in an identification text, mainly comprising a name, a place name, a mechanism name, proper noun and the like, in S4, in addition, a downstream model in S5 is a model for a downstream task based on the pre-training model, and in S4, the construction method of the whole model is that the downstream model in S4 is overlapped by the pre-training model.
Further, the BIO label in S1 is a labeling mode for labeling NER entities in the corpus by BIO, wherein PER represents a name of a person, LOC represents a place, ORG represents a mechanism, and the rest words are O.
Further, the training of the general entity extraction model and the training of the special entity extraction model in the financial field in S5 are independent of each other and can be performed synchronously.
The invention provides a method for improving the public opinion entity recognition rate in the financial risk prevention and control field through improvement, and compared with the prior art, the method has the following improvement and advantages:
(1) According to the invention, a deep learning-based method is utilized, and the extraction results of the general domain NER model based on the NLP pre-training model and the special domain NER model based on the NLP pre-training model are respectively trained and fused, so that the extraction rate of public opinion entity identification in the financial domain is improved.
(2) In the invention, in the special corpus, the BIO labeling is only carried out on financial entity abbreviations and new financial entities, and the labeling method for labeling general entities is not carried out, thereby greatly improving the training speed of the model and being beneficial to improving the accuracy.
Drawings
The invention is further explained below with reference to the drawings and examples:
fig. 1 is a schematic flow chart of the structure of the present invention.
Detailed Description
The following detailed description of the present invention will be provided with reference to fig. 1, in which the technical solutions of the embodiments of the present invention are clearly and completely described, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.
The invention provides a method for improving the identification rate of public opinion entities in the field of financial risk prevention and control through improvement, as shown in fig. 1, comprising the following steps:
s1, collecting general domain corpus, and labeling a general entity in the corpus with BIO labels;
S2, collecting corpora in the financial field, and only labeling the financial abbreviations and the special entities of new financial entities in the corpora, and not labeling the general entities;
S3, selecting a deep learning NLP pre-training model;
S4, selecting a downstream model suitable for the NER task;
S5, constructing an integral model for training NER tasks, training the integral model by using the general corpus collected in S1 to obtain a general entity extraction model, and training the integral model by using the financial domain-specific corpus collected in S2 to obtain a financial domain-specific entity extraction model;
S6, for public opinion needing to be extracted, respectively utilizing the two models obtained through training in the S5 to independently extract, wherein the general entity extraction model is responsible for extracting general entity, the new entity extraction model in the financial field is responsible for extracting entity short and new entity professional entity, and entity extraction results obtained by fusing the two models are obtained, so that the entity extraction results of the public opinion are obtained.
Further, the BIO label in S1 is a labeling mode for labeling NER entities in the corpus by BIO, wherein PER represents a name of a person, LOC represents a place, ORG represents a mechanism, and the rest words are O.
In S3, NLP is natural language processing, which is a process of processing, understanding and applying human language by a computer, in S3, a pre-training model is training a language model through a large number of unlabeled language texts to obtain a set of model parameters, the model and corresponding parameters are pre-training models, in S3, deep learning is an algorithm for carrying out characterization learning on data by taking a multi-layer artificial neural network as a framework, a common deep learning framework is provided with the deep neural network, the convolutional neural network, the cyclic neural network and the like, in S4, NER is named entity identification, which means identification of entities with specific meanings in the text, mainly comprising names, place names, organization names, proper names and the like, in S4, a downstream model is a model for a downstream task based on the pre-training model, and in S5, the construction method of the whole model is that the downstream model in S4 is overlapped by the pre-training model.
And S5, training of the general entity extraction model and the special entity extraction model in the financial field are independent from each other and can be synchronously performed.
The following are further enumerated as preferred embodiments or application examples to assist those skilled in the art in better understanding the technical content of the present invention and the technical contribution of the present invention with respect to the prior art:
Example 1
A method for improving the public opinion entity recognition rate in the financial risk prevention and control field is shown in figure 1, and comprises the following steps:
s1, collecting general domain corpus, and labeling a general entity in the corpus with BIO labels;
S2, collecting corpora in the financial field, and only labeling the financial abbreviations and the special entities of new financial entities in the corpora, and not labeling the general entities;
S3, selecting a deep learning NLP pre-training model;
S4, selecting a downstream model suitable for the NER task;
S5, constructing an integral model for training NER tasks, training the integral model by using the general corpus collected in S1 to obtain a general entity extraction model, and training the integral model by using the financial domain-specific corpus collected in S2 to obtain a financial domain-specific entity extraction model;
S6, for public opinion needing to be extracted, respectively utilizing the two models obtained through training in the S5 to independently extract, wherein the general entity extraction model is responsible for extracting general entity, the new entity extraction model in the financial field is responsible for extracting entity short and new entity professional entity, and entity extraction results obtained by fusing the two models are obtained, so that the entity extraction results of the public opinion are obtained.
The BIO label in S1 is a labeling mode for labeling NER entities in corpus by BIO mode, wherein PER represents name, LOC represents place, ORG represents mechanism, and the rest words are O; s3, NLP is natural language processing, which is a process of processing, understanding and applying human language by a computer; s3, training the language model by a large number of unlabeled language texts to obtain a set of model parameters, wherein the model and the corresponding parameters are the pre-training model; the deep learning in S3 is an algorithm for carrying out characterization learning on data by taking a multi-layer artificial neural network as a framework, and common deep learning frameworks comprise a deep neural network, a convolutional neural network, a cyclic neural network and the like; in S4, NER is named entity recognition, namely, recognizing entities with specific meanings in texts, wherein the entities mainly comprise person names, place names, mechanism names, proper nouns and the like; the downstream model in S4 is a model for a downstream task based on the pre-trained model; the construction method of the integral model in S5 is to superimpose the downstream model in S4 by using a pre-training model; and S5, training of the general entity extraction model and the special entity extraction model in the financial field are independent from each other and can be synchronously performed.
The working principle of the invention is as follows: collecting general domain corpus, carrying out BIO label marking on general entities in the corpus, wherein BIO labels are marking modes for marking NER entities in the corpus by BIO modes, wherein PER represents names, LOC represents places, ORG represents mechanisms and the rest words are O; secondly, collecting corpora in the financial field, and only labeling the financial abbreviations and the special entities of new financial entities in the corpora by BIO labels, and not labeling the general entities by BIO; thirdly, training a language model through a large number of non-marked language texts to obtain a set of model parameters, wherein the model and the corresponding parameters are pre-training models, a deep learning NLP pre-training model is selected, the deep learning is an algorithm for carrying out characterization learning on data by taking a multi-layer artificial neural network as a framework, a common deep learning framework comprises a deep neural network, a convolution neural network, a circulation neural network and the like, and the NLP is a process for processing, understanding and applying human language by a computer; selecting a downstream model suitable for NER tasks, wherein NER is named entity recognition, namely, recognizing entities with specific meanings in texts, mainly comprising names of people, places, organizations, proper nouns and the like, and the downstream model is a model for the downstream tasks based on a pre-training model; fifthly, overlapping the downstream model in the fourth step with the pre-training model to construct an integral model for training the NER task, training the integral model by using the general corpus collected in the first step to obtain a general entity extraction model, and training the integral model by using the financial domain special corpus collected in the second step to obtain a financial domain special entity extraction model, wherein the training of the general entity extraction model and the financial domain special entity extraction model are mutually independent and can be synchronously performed; and sixthly, for the public opinion needing to be extracted, respectively utilizing the two models obtained by training in the fifth step to independently extract, wherein the general entity extraction model is responsible for extracting general entity, the new entity extraction model in the financial field is responsible for extracting entity short and new entity professional entity, and the entity extraction results obtained by fusing the two models are obtained, so that the entity extraction result of the public opinion is obtained.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (3)
1. A method for improving the public opinion entity recognition rate in the financial risk prevention and control field is characterized by comprising the following steps:
s1, collecting general domain corpus, and labeling a general entity in the corpus with BIO labels;
S2, collecting corpora in the financial field, and only labeling the financial abbreviations and the special entities of new financial entities in the corpora, and not labeling the general entities;
S3, selecting a deep learning NLP pre-training model;
S4, selecting a downstream model suitable for the NER task;
S5, constructing an integral model for training NER tasks, training the integral model by using the general corpus collected in S1 to obtain a general entity extraction model, and training the integral model by using the financial domain-specific corpus collected in S2 to obtain a financial domain-specific entity extraction model;
S6, for public opinion needing to be extracted, respectively utilizing the two models obtained through training in the S5 to independently extract, wherein the general entity extraction model is responsible for extracting general entity, the new entity extraction model in the financial field is responsible for extracting entity short and new entity professional entity, and entity extraction results obtained by fusing the two models are obtained, so that entity extraction results of the public opinion are obtained;
The method comprises the steps that in S3, NLP (natural language processing) is processed, understood and applied to human language by a computer, a pre-training model in S3 is trained through a large number of unlabeled language texts to obtain a set of model parameters, the model and corresponding parameters are pre-training models, in S3, deep learning is an algorithm for carrying out characterization learning on data by taking a multi-layer artificial neural network as a framework, a common deep learning framework is provided with the deep neural network, the convolutional neural network and the cyclic neural network, in S4 NER is named entity identification, namely identifying an entity with specific meaning in the text, mainly comprising a name, a place name, a mechanism name and proper nouns, in S4, and in S5, a downstream model is a model for downstream task based on the pre-training model, and in S4, the construction method of the whole model is that the downstream model in S4 is overlapped by the pre-training model.
2. The method for improving the identification rate of public opinion entities in the field of financial risk prevention and control of claim 1, wherein the method comprises the following steps: the BIO label in the S1 is a labeling mode for labeling NER entities in the corpus by a BIO mode, wherein PER represents a name of a person, LOC represents a place, ORG represents a mechanism, and the rest words are O.
3. The method for improving the identification rate of public opinion entities in the field of financial risk prevention and control of claim 1, wherein the method comprises the following steps: and S5, training the general entity extraction model and the special entity extraction model in the financial field is independent and synchronous.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010550784.0A CN111695346B (en) | 2020-06-16 | 2020-06-16 | Method for improving public opinion entity recognition rate in financial risk prevention and control field |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010550784.0A CN111695346B (en) | 2020-06-16 | 2020-06-16 | Method for improving public opinion entity recognition rate in financial risk prevention and control field |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111695346A CN111695346A (en) | 2020-09-22 |
| CN111695346B true CN111695346B (en) | 2024-05-07 |
Family
ID=72481629
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010550784.0A Active CN111695346B (en) | 2020-06-16 | 2020-06-16 | Method for improving public opinion entity recognition rate in financial risk prevention and control field |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111695346B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112307212A (en) * | 2020-11-11 | 2021-02-02 | 上海昌投网络科技有限公司 | Public opinion delivery monitoring method for advertisement delivery |
| CN112445913B (en) * | 2020-11-25 | 2022-09-27 | 重庆邮电大学 | Financial information negative main body judgment and classification method based on big data |
| CN112395410B (en) * | 2021-01-13 | 2021-05-14 | 北京智源人工智能研究院 | Entity extraction-based industry public opinion recommendation method and device and electronic equipment |
| CN113779992A (en) * | 2021-07-19 | 2021-12-10 | 西安理工大学 | Method for realizing BcBERT-SW-BilSTM-CRF model based on vocabulary enhancement and pre-training |
| CN114757191A (en) * | 2022-03-29 | 2022-07-15 | 国网江苏省电力有限公司营销服务中心 | Electric power public opinion field named entity recognition method and system based on deep learning |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106095753A (en) * | 2016-06-07 | 2016-11-09 | 大连理工大学 | A kind of financial field based on comentropy and term credibility term recognition methods |
| CN108229806A (en) * | 2017-12-27 | 2018-06-29 | 中国银行股份有限公司 | A kind of method and system for analyzing business risk |
| CN109614550A (en) * | 2018-12-11 | 2019-04-12 | 平安科技（深圳）有限公司 | Public sentiment monitoring method, device, computer equipment and storage medium |
| CN110990525A (en) * | 2019-11-15 | 2020-04-10 | 华融融通（北京）科技有限公司 | Natural language processing-based public opinion information extraction and knowledge base generation method |
| CN111144119A (en) * | 2019-12-27 | 2020-05-12 | 北京联合大学 | Entity identification method for improving knowledge migration |
-
2020
- 2020-06-16 CN CN202010550784.0A patent/CN111695346B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106095753A (en) * | 2016-06-07 | 2016-11-09 | 大连理工大学 | A kind of financial field based on comentropy and term credibility term recognition methods |
| CN108229806A (en) * | 2017-12-27 | 2018-06-29 | 中国银行股份有限公司 | A kind of method and system for analyzing business risk |
| CN109614550A (en) * | 2018-12-11 | 2019-04-12 | 平安科技（深圳）有限公司 | Public sentiment monitoring method, device, computer equipment and storage medium |
| CN110990525A (en) * | 2019-11-15 | 2020-04-10 | 华融融通（北京）科技有限公司 | Natural language processing-based public opinion information extraction and knowledge base generation method |
| CN111144119A (en) * | 2019-12-27 | 2020-05-12 | 北京联合大学 | Entity identification method for improving knowledge migration |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111695346A (en) | 2020-09-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111695346B (en) | Method for improving public opinion entity recognition rate in financial risk prevention and control field | |
| CN113177124B (en) | Method and system for constructing knowledge graph in vertical field | |
| WO2018028077A1 (en) | Deep learning based method and device for chinese semantics analysis | |
| CN106844349B (en) | Comment spam recognition methods based on coorinated training | |
| CN108182295A (en) | A kind of Company Knowledge collection of illustrative plates attribute extraction method and system | |
| CN110807320A (en) | Short text emotion analysis method based on CNN bidirectional GRU attention mechanism | |
| CN105893582A (en) | Social network user emotion distinguishing method | |
| US20230394247A1 (en) | Human-machine collaborative conversation interaction system and method | |
| CN111046656A (en) | Text processing method and device, electronic equipment and readable storage medium | |
| CN111444704B (en) | Network safety keyword extraction method based on deep neural network | |
| CN113901208B (en) | Method for analyzing emotion tendentiousness of mid-cross language comments blended with theme characteristics | |
| CN109376250A (en) | Entity relationship based on intensified learning combines abstracting method | |
| CN110297986A (en) | A kind of Sentiment orientation analysis method of hot microblog topic | |
| CN109446523A (en) | Entity attribute extraction model based on BiLSTM and condition random field | |
| CN112380868A (en) | Petition-purpose multi-classification device based on event triples and method thereof | |
| CN111091002B (en) | Chinese named entity recognition method | |
| CN116341519A (en) | Event causal relation extraction method, device and storage medium based on background knowledge | |
| CN106897274B (en) | Cross-language comment replying method | |
| Lin et al. | Multi-modal feature fusion with feature attention for VATEX captioning challenge 2020 | |
| EP2605150A1 (en) | Method for identifying the named entity that corresponds to an owner of a web page | |
| CN115017271B (en) | Method and system for intelligently generating RPA flow component block | |
| Yigzaw et al. | A Generic Approach towards Amharic Sign Language Recognition | |
| CN109582925A (en) | A kind of corpus labeling method and system of man-computer cooperation | |
| Tazalli et al. | Computer vision-based Bengali sign language to text generation | |
| CN113191160A (en) | Emotion analysis method for knowledge perception |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: No. 08, 27th Floor, Building C3, Jiangling North Road and Zongsan Road, Nansha District, Guangzhou City, Guangdong Province, 511455 Patentee after: Guangzhou financial technology Co.,Ltd. Country or region after: China Address before: J41, 11th Floor, 1101, Nansha Financial Building, 171 Haibin Road, Nansha District, Guangzhou City, Guangdong Province, 511455 Patentee before: Guangzhou Commodity Clearing Center Co.,Ltd. Country or region before: China |