Unlock instant, AI-driven research and patent intelligence for your innovation.
End-to-end aspect level sentiment analysis method combined with reconstructed syntactic information
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
A sentiment analysis, end-aspect-level technology, applied in the field of end-to-end aspect-level sentiment analysis, can solve the problems of not considering the influence of emotional polarity, losing long-distance dependencies, etc.
Active Publication Date: 2021-06-08
KUNMING UNIV OF SCI & TECH
View PDF8 Cites 4 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
However, the above model only considers the aspect words themselves, and does not take into account the influence of opinion words in comment sentences on the emotional polarity of aspect words. The judgment of emotional polarity is based on the current aspect word information, which will lose the long-distance useful for judging emotions. reliance on
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 more
Image
Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0033] Example 1: Such as Figure 1-2 As shown, an end-to-end-to-end emotional analysis method of combining the reconstructed sentence information, the specific steps of the method are as follows:
[0034] Step 1, pass the text code to the text encoding by the Bert pre-trained model, and obtain the word vector characterization with context characterization;
[0035] As a further scheme of the present invention, in the step 1, the BERT embedding layer uses the sentence as input, using the information of the entire sentence, and the word vector characterization is characterized by formula h. l = TRANSFORMER l (H l-1 ) Calculate, where H l The characteristics of the table L layer.
[0036] Step 2, use the AllennLP tool to design the BIAffine model, use the double emulation model Biaffine to get the initial syntax tree;
[0037] Further, the AllennLP tool used in step 2 is a depth learning model for building natural languageprocessing, which is built on Pytorch, providing advanced abs...
Embodiment 2
[0057] Example 2: Such as Figure 1-2 As shown, an end-to-end level emotional analysis method of combining the reconstructed sentence information is
[0058] Step 1: Enter the text to the BERT layer encoding. Enter a specific area comment text W n , Expressed as w = (w 1 , ..., w N ), Where n is the length of the sentence. Then package the input feature to h 0 = (X 1 , ..., x T ), Where X n (t∈ [1, n]) is input W n Corresponding words embedding, position embedding, and segment embedded combinations;
[0059] Step 2: Refine the word characteristics. Through the steps, you get the packaged text, refine the word class by layer by L-layer transformerlayer, calculate the context representation of W corresponding to W DIM h Indicates the dimension representing the vector.
[0060] The characteristics of the first layer represent the calculation as follows:
[0062] Step 3: Predict the label, will h l As a context of the input word, and use them t...
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 relates to an end-to-end aspect level sentiment analysis method in combination with reconstructed syntactic information, and belongs to the technical field of natural languageprocessing. The method comprises the following steps: encoding a text through a Bert pre-training model to obtain word vector representation with context representation; the method comprises the following steps: obtaining an initial syntactic tree by using a double affine model Biaffine; remodeling and pruning the initial syntactic tree through syntactic rules to obtain a new dependency tree; encoding the new dependency tree by using an attention neural network to obtain reconstructed syntactic features; performing feature splicing fusion by using the obtained word vector representation and the reconstructed syntactic features, and inputting the fused features to a downstream sequence labeling model; and obtaining an aspect-level sentiment analysis result through an output result of the downstream sequence labeling model. According to the method, the effective syntactic dependency relationship between the aspect word and the viewpoint word can be obtained, and the performance of an aspect-level sentiment analysis task is improved.
Description
Technical field [0001] The present invention relates to an end-to-end-ended emotional analysis method of combining the reconstructor information, which is a natural languageprocessing technology. Background technique [0002] Traditionally, aspective emotional analysis (ABSA) tasks can be divided into two sub-tasks, namely, the word extraction tasks and aspect emotional analysis tasks. The purpose of the terms of words is to detect the viewing objectives mentioned in commentary, and conduct extensive research on them. The purpose of aspect emotion classification is to find the corresponding opinions through the territory to help model predict the emotional polarity of the given object. At present, most of the work of ABSA is intended to address one of the sub-tasks. In order to apply these existing methods to the actual environment, that is, the evaluation object is not only extracted, but it is predicted that its emotional polarity is predicted, a typical method is to connect t...
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
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to View More 
Login to View More