Aspect-level sentiment analysis for based on joint aspect and position hierarchy attention mechanism network

2021 ◽  
pp. 1-12
Author(s):  
Dangguo Shao ◽  
Qing An ◽  
Kun Huang ◽  
Yan Xiang ◽  
Lei Ma ◽  
...  
Keyword(s):  
Information Retrieval ◽  
Sentiment Analysis ◽  
Attention Mechanism ◽  
Sentiment Classification ◽  
The Other ◽  
Position Information ◽  
Vector Information ◽  
Feature Information ◽  
Joint Approach ◽  
The One

The purpose of aspect-level sentiment analysis is to identify the contextual sentence expressions given by sentiment for some aspects. For previous works, many scholars have proved the importance of the interaction between aspects and contexts. However, most existing methods ignore or do not specifically capture the position information of the aspect targets in the sentence. Thus, we propose an aspect-level sentiment analysis based on joint aspect and position hierarchy attention mechanism network. At the same time, the model adopts a joint approach to make the model of the aspect features and the position features. On the one hand, this method clearly captures the interaction between aspect words and context when inputting word vector information. On the other hand, this method can enhance the importance of position information in the sentence and boost the information retrieval ability of the model. Additionally, the model utilizes a hierarchical attention mechanism to extract feature information and to differentiate sentiment towards, which is similar to filtering useless information again. Experiment on the SemEval 2014 dataset represent that our model achieves better performance on aspect-level sentiment classification.

scholarly journals Aspect-Level Sentiment Analysis Based on Position Features Using Multilevel Interactive Bidirectional GRU and Attention Mechanism

2020 ◽  
Vol 2020 ◽  
pp. 1-13
Author(s):  
Xiaodi Wang ◽  
Xiaoliang Chen ◽  
Mingwei Tang ◽  
Tian Yang ◽  
Zhen Wang
Keyword(s):  
Neural Network ◽  
Sentiment Analysis ◽  
Network Models ◽  
Attention Mechanism ◽  
Sentiment Classification ◽  
Classification Model ◽  
Neural Network Models ◽  
Position Information ◽  
Gated Recurrent Unit ◽  
Relative Position Information

The aim of aspect-level sentiment analysis is to identify the sentiment polarity of a given target term in sentences. Existing neural network models provide a useful account of how to judge the polarity. However, context relative position information for the target terms is adversely ignored under the limitation of training datasets. Considering position features between words into the models can improve the accuracy of sentiment classification. Hence, this study proposes an improved classification model by combining multilevel interactive bidirectional Gated Recurrent Unit (GRU), attention mechanisms, and position features (MI-biGRU). Firstly, the position features of words in a sentence are initialized to enrich word embedding. Secondly, the approach extracts the features of target terms and context by using a well-constructed multilevel interactive bidirectional neural network. Thirdly, an attention mechanism is introduced so that the model can pay greater attention to those words that are important for sentiment analysis. Finally, four classic sentiment classification datasets are used to deal with aspect-level tasks. Experimental results indicate that there is a correlation between the multilevel interactive attention network and the position features. MI-biGRU can obviously improve the performance of classification.

Sentiment Classification

2015 ◽  
pp. 1-26
Author(s):  
Jalel Akaichi
Keyword(s):  
Support Vector Machine ◽  
Text Mining ◽  
Sentiment Analysis ◽  
Training Model ◽  
Sentiment Classification ◽  
The Other ◽  
Support Vector ◽  
Analysis Techniques ◽  
Sentiment Lexicon ◽  
And Behavior

In this work, we focus on the application of text mining and sentiment analysis techniques for analyzing Tunisian users' statuses updates on Facebook. We aim to extract useful information, about their sentiment and behavior, especially during the “Arabic spring” era. To achieve this task, we describe a method for sentiment analysis using Support Vector Machine and Naïve Bayes algorithms, and applying a combination of more than two features. The output of this work consists, on one hand, on the construction of a sentiment lexicon based on the Emoticons and Acronyms' lexicons that we developed based on the extracted statuses updates; and on the other hand, it consists on the realization of detailed comparative experiments between the above algorithms by creating a training model for sentiment classification.

scholarly journals A SURVEY ON BUSINESS FACILITIES ANALYSIS

2017 ◽  
Vol 10 (13) ◽  
pp. 15
Author(s):  
S Joshika
Keyword(s):  
Sentiment Analysis ◽  
Sentiment Classification ◽  
The Other ◽  
Classification Algorithms ◽  
Data Set ◽  
Classification Techniques ◽  
Survey Papers ◽  
A Site

Yelp connects people to great local businesses in USA which maintains a site to search and find any business in USA. This helps user to compare the businesses based on the star ratings and reviews given by other users to identify the best company among the available according to their need. The data-set provided in Yelp challenge contains tip, review, users, check-in, and business details which is shortly called as TURBO set was used by the participants in various ways to find interesting patterns. This paper focuses various surveys made on pre-processing; sentiment analysis; sentiment classification techniques and various classification algorithms proposed that results better performance than the other existing algorithms. The survey papers have mostly applied the algorithms on yelp data-set and other papers have applied on different data’s. 

scholarly journals A Span-based Joint Model for Opinion Target Extraction and Target Sentiment Classification

2019 ◽  
Author(s):  
Yan Zhou ◽  
Longtao Huang ◽  
Tao Guo ◽  
Jizhong Han ◽  
Songlin Hu
Keyword(s):  
Sentiment Analysis ◽  
State Of The Art ◽  
Joint Model ◽  
The State ◽  
Attention Mechanism ◽  
Sentiment Classification ◽  
Global Information ◽  
Target Extraction ◽  
Benchmark Datasets ◽  
Tagging Methods

Target-Based Sentiment Analysis aims at extracting opinion targets and classifying the sentiment polarities expressed on each target. Recently, token based sequence tagging methods have been successfully applied to jointly solve the two tasks, which aims to predict a tag for each token. Since they do not treat a target containing several words as a whole, it might be difficult to make use of the global information to identify that opinion target, leading to incorrect extraction. Independently predicting the sentiment for each token may also lead to sentiment inconsistency for different words in an opinion target. In this paper, inspired by span-based methods in NLP, we propose a simple and effective joint model to conduct extraction and classification at span level rather than token level. Our model first emulates spans with one or more tokens and learns their representation based on the tokens inside. And then, a span-aware attention mechanism is designed to compute the sentiment information towards each span. Extensive experiments on three benchmark datasets show that our model consistently outperforms the state-of-the-art methods.

Negative emotion diffusion and intervention countermeasures of social networks based on deep learning

2020 ◽  
Vol 39 (4) ◽  
pp. 4935-4945
Author(s):  
Qiuyun Cheng ◽  
Yun Ke ◽  
Ahmed Abdelmouty
Keyword(s):  
Deep Learning ◽  
Sentiment Analysis ◽  
Negative Emotion ◽  
Sentiment Classification ◽  
Classification Model ◽  
Learning Models ◽  
Data Set ◽  
Position Information ◽  
Processing Technologies ◽  
And Training

Aiming at the limitation of using only word features in traditional deep learning sentiment classification, this paper combines topic features with deep learning models to build a topic-fused deep learning sentiment classification model. The model can fuse topic features to obtain high-quality high-level text features. Experiments show that in binary sentiment classification, the highest classification accuracy of the model can reach more than 90%, which is higher than that of commonly used deep learning models. This paper focuses on the combination of deep neural networks and emerging text processing technologies, and improves and perfects them from two aspects of model architecture and training methods, and designs an efficient deep network sentiment analysis model. A CNN (Convolutional Neural Network) model based on polymorphism is proposed. The model constructs the CNN input matrix by combining the word vector information of the text, the emotion information of the words, and the position information of the words, and adjusts the importance of different feature information in the training process by means of weight control. The multi-objective sample data set is used to verify the effectiveness of the proposed model in the sentiment analysis task of related objects from the classification effect and training performance.

Using a sentiment analysis for the examination of tourism blogs – a step by step methodological reflection process

2021 ◽  
Vol 13 (2) ◽  
pp. 167-190
Author(s):  
Gero Szepannek ◽  
Laila Westphal ◽  
Werner Gronau ◽  
Tine Lehmann
Keyword(s):  
Content Analysis ◽  
Sentiment Analysis ◽  
Computational Methods ◽  
Research Question ◽  
The Other ◽  
Data Sets ◽  
New Approach ◽  
Other Hand ◽  
The One ◽  
The Given

Abstract The article at hand is driven by a methodological interest in the opportunities and challenges of applying an automated text mining approach, particularly a sentiment analysis on various tourism blogs at the same time. The study aims to answer the question to what extent advanced computational methods can improve the data acquisition and analysis of unstructured data sets stemming from various blogs and forums. Furthermore, the authors intend to explore to what extent the sentiment analysis is able to objectify the qualitative results identified by an earlier analysis by the authors using content analysis done by thematic coding. For the purpose of the specific tourism research question in this paper a new approach is proposed, which consists of a combination of sentiment analyses, supervised learning, and dimensionality reduction in order to identify terms that strongly load on specific emotions. The contribution indicates on the one hand, that advanced computational methods have their own specific constraints, but on the other hand, are able to provide a richer and deeper analysis following a quantitative approach. Several issues have to be taken into account, such as data protection constraints, the need for data cleaning, such as word stemming, dimension reduction, such as removal of custom stop words, and the development of descent ontologies. On the other hand, the quantitative method also provides, due to its standardised procedure, a less subjective insight in the given content, but is not less time consuming than traditional content analysis.

scholarly journals System & Contrast

2016 ◽  
Vol 33 (5) ◽  
pp. 631-661 ◽  
Author(s):  
Frédéric Bimbot ◽  
Emmanuel Deruty ◽  
Gabriel Sargent ◽  
Emmanuel Vincent
Keyword(s):  
Information Retrieval ◽  
Data Processing ◽  
Music Perception ◽  
Minimum Description Length ◽  
The Other ◽  
Carrier System ◽  
Structural Annotation ◽  
Other Hand ◽  
The One

This article introduces the System &Contrast (S&C) model, which aims at describing the inner organization of structural segments within music pieces as: (i) a carrier system, i.e., a sequence of morphological elements forming a network of self-deducible syntagmatic relationships, and (ii) a contrast, i.e., a substitutive element, usually the last one, which departs from the logic implied by the carrier system. Initially used for the structural annotation of pop songs (Bimbot, Deruty, Sargent, & Vincent, 2012), the S&C model provides a framework to describe implication patterns in musical segments by encoding similarities and relations between its elements. It is applicable at several timescales to various musical dimensions in a polymorphous way, thus offering an attractive meta-description of musical contents. We formalize the S&C model, illustrate how it applies to music and establish its filiation with Narmour’s implication-realization model (Narmour, 1990, 1992) and cognitive rule-mapping (Narmour, 2000). We introduce the minimum description length scheme as a productive paradigm to support the estimation of S&C descriptions. The S&C model highlights promising connections between music data processing and information retrieval on the one hand, and modern theories in music perception, cognition and semiotics on the other hand, together with interesting perspectives in Musicology.

scholarly journals Hybrid CNN Classification for Sentiment Analysis under Deep Learning

2020 ◽  
Vol 9 (5) ◽  
pp. 1473-1480
Keyword(s):  
Sentiment Analysis ◽  
Language Processing ◽  
Sentiment Classification ◽  
The Other ◽  
Complex Sentence ◽  
Abstract Noun ◽  
Complex Sentences ◽  
Syntactic Features ◽  
Twitter Data ◽  
Abstract Nouns

Sentiment Analysis (SA) is a popular field in Natural Language Processing (NLP) which focuses on the human emotions by analyzing the lexical and syntactic features. This paper presents an efficient method to find and extract the strong emotions for the sentiment classification using the proposed hybrid Convolutional Neural Networks - Global Vectors - Complex Sentence Searching - ABstract Noun Searching (CNN-GloVe-CSS-ABNS) model. The strong emotions are mostly found in the abstract nouns than the adjectives and adverbs present in the sentences. This research aims in extracting the complex sentences with abstract nouns for the sentiment classification from the twitter data. To extract the complex sentences, the proposed Complex Sentence Searching (CSS) algorithm was used. On the other hand, another proposed algorithm named, ABstract Noun Searching (ABNS) algorithm was used for identifying the abstract nouns in the sentences based on their position in the sentences. The results of this study presents that the proposed CNN-GloVe-CSS-ABNS model outperforms the other proposed models as well as the existing models, by producing an of accuracy 94.87 per cent in sentiment classification.

scholarly journals Stance Classification with Target-specific Neural Attention

2017 ◽  
Author(s):  
Jiachen Du ◽  
Ruifeng Xu ◽  
Yulan He ◽  
Lin Gui
Keyword(s):  
Neural Network ◽  
Sentiment Analysis ◽  
State Of The Art ◽  
Attention Mechanism ◽  
Sentiment Classification ◽  
Specific Information ◽  
Target Information ◽  
Specific Target ◽  
Proposed Model ◽  
Text Content

Stance classification, which aims at detecting the stance expressed in text towards a specific target, is an emerging problem in sentiment analysis. A major difference between stance classification and traditional aspect-level sentiment classification is that the identification of stance is dependent on target which might not be explicitly mentioned in text. This indicates that apart from text content, the target information is important to stance detection. To this end, we propose a neural network-based model, which incorporates target-specific information into stance classification by following a novel attention mechanism. In specific, the attention mechanism is expected to locate the critical parts of text which are related to target. Our evaluations on both the English and Chinese Stance Detection datasets show that the proposed model achieves the state-of-the-art performance.

scholarly journals Compilation and evaluation of the Spanish SatiCorpus 2021 for satire identification using linguistic features and transformers

2022 ◽  
Author(s):  
José Antonio García-Díaz ◽  
Rafael Valencia-García
Keyword(s):  
Social Media ◽  
Information Retrieval ◽  
The Other ◽  
Linguistic Features ◽  
Feature Sets ◽  
Retrieval Systems ◽  
Other Hand ◽  
Extensive Evaluation ◽  
Information Retrieval Systems ◽  
The One

AbstractSatirical content on social media is hard to distinguish from real news, misinformation, hoaxes or propaganda when there are no clues as to which medium these news were originally written in. It is important, therefore, to provide Information Retrieval systems with mechanisms to identify which results are legitimate and which ones are misleading. Our contribution for satire identification is twofold. On the one hand, we release the Spanish SatiCorpus 2021, a balanced dataset that contains satirical and non-satirical documents. On the other hand, we conduct an extensive evaluation of this dataset with linguistic features and embedding-based features. All feature sets are evaluated separately and combined using different strategies. Our best result is achieved with a combination of the linguistic features and BERT with an accuracy of 97.405%. Besides, we compare our proposal with existing datasets in Spanish regarding satire and irony.

Sign in / Sign up

Export Citation Format

Share Document