Viewpoint mining method and device based on reading understanding
A technology of reading comprehension and opinion mining, applied in the fields of electronic digital data processing, natural language data processing, special data processing applications, etc., to achieve the effect of accurate opinion mining results
Active Publication Date: 2021-08-27
NAT COMP NETWORK & INFORMATION SECURITY MANAGEMENT CENT +1
View PDF5 Cites 3 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
In addition, the classification of emotion and stance in previous models is an independent process, but there is often an implicit synergistic relationship between emotion and stance
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0041] Below in conjunction with embodiment and accompanying drawing, the present invention will be described in further detail, so that those skilled in the art can implement according to referring to description text.
[0042] It should be understood that terms such as "having", "comprising" and "including" used herein do not exclude the presence or addition of one or more other elements or combinations thereof.
[0043] Explanation of terms:
[0044] Mask: word masking. Replace specified words in the text with the generic masking labels [MASK] defined in BERT.
[0045]BERT: A general-purpose NLP deep learning model proposed by goole, which can be used for various NLP tasks after fine-tuning. See Jacob Devlin, Ming-Wei Chang, Kenton Lee, Kristina Toutanova: BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. NAACL-HLT(1) 2019:4171-4186.
[0046] Embedding: The natural original data is a low-dimensional manifold embedded in the high-dimensiona...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a viewpoint mining method based on reading understanding, and the method comprises the steps: constructing a domain sentiment viewpoint knowledge base which comprises domain sentiment viewpoint words, and each domain sentiment viewpoint word corresponds to a sentiment classification tag and a vertical classification tag; based on the domain emotion viewpoint knowledge base and the event viewpoint training text set, training a pre-training language model, obtaining an emotion pre-training language model, and embeding emotion and viewpoint information representing an input text in the emotion pre-training language model; extracting an event viewpoint text from the event text to be extracted; inputting the event viewpoint text into the emotion pre-training language model, and performing encoding, sentence feature extraction and classification on the output content to obtain the emotion and the standing site of the viewpoint in the event text to be extracted. The invention also discloses a viewpoint mining device based on reading understanding. The method has the advantage that the viewpoint mining result is more accurate.
Description
technical field [0001] The invention relates to the field of natural language processing. More specifically, the present invention relates to a method and device for mining ideas based on reading comprehension. Background technique [0002] Opinion mining includes two steps: identify and classify. The main method of existing opinion extraction research is the rule-based model. This type of method designs special trigger words according to the text characteristics and actual needs of the event, and combines named entity recognition technology to identify key information such as names of people and organizations, and design opinion Extraction rules extract expert opinions in events. However, this type of model needs to design extraction rules according to actual needs. It is an unsupervised model, lacks an understanding of the semantic information of the overall event document, and cannot understand the directional evaluation objects, specific emotions, and specific position...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/35G06F40/289G06F40/30G06K9/62G06F40/216G06N5/02
CPCG06F16/35G06F40/289G06F40/30G06F40/216G06N5/022G06F18/22
Inventor 陈志鹏刘春阳张旭王鹏董善为李倩虞世城李建欣
Owner NAT COMP NETWORK & INFORMATION SECURITY MANAGEMENT CENT