Chinese comments sentiment classification based on word2vec and SVMperf.

Dongwen Zhang,Hua Xu,Zengcai Su,Yunfeng Xu

DOI: https://doi.org/10.1016/j.eswa.2014.09.011

2015-01-01

Abstract:We achieve similar features clustering using word2vec.A method for sentiment classification based on word2vec and SVMperf is proposed.Word2vec can extract deep semantic features between words.SVMpe...

What problem does this paper attempt to address?

Chinese Sentiment Classification Using A Neural Network Tool-Word2vec

Zengcai Su,Hua Xu,Dongwen Zhang,Yunfeng Xu

DOI: https://doi.org/10.1109/MFI.2014.6997687

2014-01-01

Abstract:Sentiment classification is the main and popular task in the field of sentiment analysis. Most of the existing researches focus on how to extract the effective features, such as lexical features and syntactic features, while limited work has been done on the extraction of semantic features, which can make more contributions to sentiment classification. This paper presents a method for sentiment classification based on word2vec. Word2vec is a tool, which establishes the neural network models to learn the vector representations of words in the high dimensional vector space. So it can extract the deep semantic relationships between words. In this paper, firstly, we cluster the similar features together using word2vec. And then we use word2vec again to learn the word representations as candidate feature vectors. After feature selection, the SVMperf package is adopted to train and classify the comment texts. To conduct the experiments, we collect a large number of Chinese comments on clothing products as data set. The experimental results show that the accuracy of sentiment classification is over 90 percent, which proves the effectiveness of proposed method for Chinese sentiment classification.
Sentiment Classification for Chinese Reviews: a Comparison Between SVM and Semantic Approaches

Q Ye,B Lin,YJ Li

DOI: https://doi.org/10.1109/icmlc.2005.1527335

2005-01-01

Abstract:Web content mining is intended to help people to discover valuable information from large amount of unstructured data on the Web. Sentiment classification aims to mining the Web content of product reviews by classifying the reviews into positive or negative opinions. Such kind of classification approaches could help both consumers and sellers in making their decisions. But it is also a complicated task with great challenge. This paper conducted a comparison between the SVM approach and semantic approach for sentiment classification of Chinese reviews and also proposed some improvement for sentiment classification approaches. Experimental result indicated that, compared with previous researches for English reviews, the performance of both approaches for Chinese reviews sentiment classification are acceptable, while the support vector machine approach has better performance than the semantic orientation approach.
Chinese Micro-Blog Emotion Classification By Exploiting Linguistic Features And Svmperf

Hua Xu,Fan Zhang,Jiushuo Wang,Weiwei Yang

DOI: https://doi.org/10.1007/978-3-319-30319-2_10

2016-01-01

Abstract:These years, micro-blog emotion mining becomes one of the research hotspots in social network data mining. Different from state of the art study, this paper presents a novel method for emotion classification, which is SVMperf based method combined with syntactic structure of Chinese micro-blogs. The classified emotion type includes Happiness, Anger, Disgust, Fear, Sadness and Surprise. For the proposed method, an emotional lexicon is constructed and linguistic features are extracted from micro-blog corpus firstly. Secondly, for the current feature space dimension is higher, Chi-square test is used to extract the high-frequency and high-class relevance keywords. At the same time, Pointwise Mutual Information (PMI) is used to pick the effective low frequency words in feature dimension reduction, which can reduce the computational complexity. Finally, SVMperf is applied for the emotion classification. In order to illustrate the effectiveness of the algorithm, LIBSVM and SVM-Light are used as the baseline. The data from Sina Micro-blog (weibo.com) have been used as the experiment data. The experiment results demonstrate that all the above features contribute to emotion classification inmicro-blogs, and the results validate the feasibility of the proposed approach. It also shows that SVMperf is an appropriate choice of classifier for emotion classification.
A Multi-dimension and Multi-granularity Feature Fusion Method for Chinese Microblog Sentiment Classification

Yanxia Chen,Tingting Wei,Leyao Qu,Xiaoyue Liu

DOI: https://doi.org/10.1145/3639479.3639490

2023-12-27

Abstract:Chinese microblog comments exhibit a wide range of expressions, including brief terms and internet slang. Traditional single-feature models may not capturing this diversity comprehensively and in effectively discerning the usage of sentiment words, language structure, and contextual information. Consequently, they tend to exhibit suboptimal performance, particularly when dealing with informally expressed datasets in this domain. This study introduces a novel approach for Chinese microblog sentiment classification that leverages multi-dimensional (common text features and sentiment features) and multi-granularity (word-level and character-level) sentiment features. To achieve this, we first employ word2vec to train separate word-level and character-level embedding vectors. And then, we introduce LSTM (Long Short-Term Memory) neural networks to capture sentence-level representations. Additionally, self-attention mechanisms are incorporated to assess the importance of sentiment words at both word and character levels, finally integrating them into the LSTM network to capture the sentimental features that incorporate contextual semantics. We conduct experiments on two different datasets. Our experimental results reveal that our method performs comparably to competing approaches on high-quality datasets. However, on lower-quality datasets, it outperforms other methods, demonstrating the robustness of our proposed approach. Our innovative method provides an effective solution to address the challenges of Chinese microblog sentiment classification, with wide-ranging potential for practical applications.

Computer Science
Sentiment Text Classification of Customers Reviews on the Web Based on SVM

Huosong Xia,Min Tao,Yi Wang

DOI: https://doi.org/10.1109/icnc.2010.5584077

2010-01-01

Abstract:As a developing endeavor of data mining on semi-structured information, sentiment analysis to the comments on the Internet has aroused people's great interest recently. This paper analysis the influence of different stop word removal methods on the result of text classification and represent the more effective stop word removal list. The experiment bases on the sentiment comments which have been grasped on the Web, using two different kinds of feature selection, choose the TF-IDF function to calculate the feature weights. Implement the classification with the technology of SVM.
Chinese Micro-Blog Emotion Classification by Exploiting Linguistic Features and SVM $$^{\Textit{perf}}$$

Hua Xu,Fan Zhang,Jiushuo Wang,Weiwei Yang

DOI: https://doi.org/10.1007/978-3-319-30319-2_10

2016-01-01

Abstract:These years, micro-blogMicro-blog emotion mining becomes one of the research hotspots in social network data mining. Different from state of the art study, this paper presents a novel method for emotion classificationEmotion classification , which is SVM basedSVM method combined with syntactic structure of Chinese micro-blogs. The classified emotion type includes Happiness, Anger, Disgust, Fear, Sadness and Surprise. For the proposed method, an emotional lexicon is constructed and linguistic features are extracted from micro-blog corpus firstly. Secondly, for the current feature space dimension is higher, Chi-square test is used to extract the high-frequency and high-class relevance keywords. At the same time, Pointwise Mutual Information (PMI) is used to pick the effective low frequency words in feature dimension reduction, which can reduce the computational complexity. Finally, SVM $$^{\textit{perf}}$$ is applied for the emotion classification. In order to illustrate the effectiveness of the algorithm, LIBSVM and SVM-Light are used as the baseline. The data from Sina Micro-blog (weibo.com) have been used as the experiment data. The experiment results demonstrate that all the above features contribute to emotion classification in micro-blogs, and the results validate the feasibility of the proposed approach. It also shows that SVM SVM is an appropriate choice of classifier for emotion classification.
Combining Vector Space Features and Convolution Neural Network for Text Sentiment Analysis.

Yun Wang,Xu An Wang,Jindan Zhang,Chenghai Yu

DOI: https://doi.org/10.1007/978-3-319-93659-8_71

2019-01-01

Abstract:With the advanced tendency of big data, there are billions of text messages posted on the Internet to express peoples' views and opinions all the time. And with the explosive growth of text information flow in social network, the topic detection and the sentiment analysis have become important researches in the field of NLP (natural language processing). There are two ways to analyze the sentiment tendency: traditional statistics method and machine learning method. At present, with the extensive application of machine learning method in NLP, various neural network models have achieved commendable results in the research of sentiment classification. However, such methods require a long period in training and learning process, and the training effect may be over fitting depended on the training set. In addition, such methods usually do not consider the vector space characteristics and lacks of the use of the word sentiment label. Standing around problem, this paper creates a sentiment classification method which combined with the feature of text vector space and the convolutional neural network method. Firstly, the words are ranked and selected according to the spatial distribution features in text information. Then the processed words are mapped into abstract vectors based on the existing dictionary resources. And the convolutional neural network is utilized to extract features of abstract vectors for the sentiment classification. This paper puts forward the relevant methods obtained superb performance in the Chinese tendency analysis evaluation (COAE 2014) data sets.
SVM-Based Sentiment Analysis Algorithm of Chinese Microblog under Complex Sentence Pattern

Jundong Zhang,Chenglin Zhao,Fangmin Xu,Peiying Zhang

DOI: https://doi.org/10.1007/978-981-10-3229-5_86

2016-01-01

Abstract:With the development of Web2.0 era, as local information publishing and social networking platform of Twitter, microblog has become an important medium for people to share and propagate information. Sentiment classification for microblog has also become research hotspot in natural language processing field. By analyzing existing sentiment classification features and complex sentence patterns of microblog and directing at defects of current microblog sentiment classification in feature selection and extraction, this paper combined semantic relation between complex sentences and sentence features of complete sentence based on proposing features of sentence-level fine-grained embedding features and semantic features under complex sentence pattern so as to conduct effective analysis of microblog sentiment features under complex sentence context. It used SVM classification model to conduct comparative experiment, and results indicated that feature selection method proposed in this paper could improve performance of microblog sentiment analysis.
An Empirical Study on Sentiment Classification of Chinese Review using Word Embedding

Yiou Lin,Hang Lei,Jia Wu,Xiaoyu Li

DOI: https://doi.org/10.48550/arXiv.1511.01665

2015-11-05

Abstract:In this article, how word embeddings can be used as features in Chinese sentiment classification is presented. Firstly, a Chinese opinion corpus is built with a million comments from hotel review websites. Then the word embeddings which represent each comment are used as input in different machine learning methods for sentiment classification, including SVM, Logistic Regression, Convolutional Neural Network (CNN) and ensemble methods. These methods get better performance compared with N-gram models using Naive Bayes (NB) and Maximum Entropy (ME). Finally, a combination of machine learning methods is proposed which presents an outstanding performance in precision, recall and F1 score. After selecting the most useful methods to construct the combinational model and testing over the corpus, the final F1 score is 0.920.

Computation and Language
Combining convolutional neural network and support vector machine for sentiment classification

Cao Yuhui,Xu Ruifeng,Chen Tao

DOI: https://doi.org/10.1007/978-981-10-0080-5_13

2015-01-01

Abstract:In recent years, the classifiers based on convolutional neural network (CNN) and word embedding achieved good performances in sentiment classification tasks. However, the CNN-based model simply uses a fully connected layer for classification and it cannot perform a nonlinear classification efficiently compared to the support vector machine (SVM) classifier. Target to this problem, in this paper, we combine CNN and SVM for sentiment classification. Firstly, continuous bag of word (CBOW) model is applied to construct word embedding. CNN is then utilized to learn feature vector representation corresponding to each sentence. The learned vector representations are fed to a SVM classifier as features for sentiment classification. Evaluations on the NLPCC2014 Sentiment Classification with Deep Learning Technology Task datasets (in short, NLPCC-SCDL) show that our model outperforms the top system in the NLPCC 2014 evaluation, on both English and Chinese sides. © Springer Science+Business Media Singapore 2015.
Sentiment Classification for Chinese News Using Machine Learning Methods

XU Jun,DING Yu-xin,WANG Xiao-long

DOI: https://doi.org/10.3969/j.issn.1003-0077.2007.06.013

2007-01-01

Abstract:In this paper,we study how to apply machine learning techniques to solve sentiment classification problems.The main task of sentiment classification is to determine whether news or reviews is negative or positive.Naive Bayes and Maximum Entropy classification are used for the sentiment classification of Chinese news and reviews.The experimental results show that the methods we employed perform well.The accuracy of classification can achieve about 90%.Moreover,we find that selecting the words with polarity as features,negation tagging and representing test documents as feature presence vectors can improve the performance of sentiment classification.Conclusively,sentiment classification is a more challenging problem.
Implementing Chinese new word discovery and POS tagging based on support vector machine

Yuejie Zhang,Hui Yang,Tao Zhang

2009-01-01

Journal of Computational Information Systems

Abstract:This paper proposes a SVM-based unified mechanism for Chinese New Word Discovery and POS Tagging. Both New Word Discovery and POS Tagging are defined as a binary classification problem, in which many previous morphological features are considered. Some new features are introduced, such as affix information and context information, as well as some constraints. Some rules are also added to improve the identification performance. The experimental results show that these features, constraints and rules are all useful for New Word Discovery. The F-measure is 64.58%, which is 6.82% higher than the predefined baseline. Based on the similar processing pattern, the precision of 90.81% for POS Tagging of new words can also be achieved. 1553-9105/ Copyright © 2009 Binary Information Press.
Sentiment Classification for Consumer Word-of-Mouth in Chinese: Comparison Between Supervised and Unsupervised Approaches

Ziqing Zhang,Qiang Ye,Wenying Zheng,Yijun Li

DOI: https://doi.org/10.2991/icebi.2010.56

2010-01-01

Abstract:Sentiment classification aims at mining word-of-mouth, reviews of consumers, for a product or service by automatically classifying reviews as positive or negative.Few empirical studies have been conducted in comparing the different effects between machine learning and semantic orientation approaches on Chinese sentiment classification.This paper adopts three supervised learning approaches and a web-based semantic orientation approach, PMI-IR, to Chinese reviews.The results show that SVM outperforms naive bayes and N-gram model on various sizes of training examples, but does not obviously exceeds the semantic orientation approach when the number of training examples is smaller than 300.
Feature representation based on sentimental orientation classification

Gongshen Liu,Wenlei He,Jie Zhu,Huoyao Lai

2011-01-01

China Communications

Abstract:Online reviews and comments are important information resources for people. A new model, called Sentiment Vector Space Model (SVSM), for feature selection and weighting is proposed to predict the sentiment orientation of comments and reviews, e.g., sorting out positive reviews from negative ones. Different from that of topic-oriented classification, feature selection of sentiment orientation prediction focuses on language characteristics. Different from traditional algorithms for sentiment classification, this model integrates grammatical knowledge and takes topic correlations into account. Features are extracted, and the similarity between these features and the topic are also computed. The feature similarity is taken as a factor when evaluating the polarity of opinions. The experimental results show that the proposed model is more effective in identifying sentiment orientation than most of the traditional techniques.
Chinese text classification based on character-level CNN and SVM

Huaiguang Wu,Daiyi Li,Ming Cheng

DOI: https://doi.org/10.1504/IJIIDS.2019.102940

2019-11-25

International Journal of Intelligent Information and Database Systems

Abstract:Aiming at the problems of curse of dimensionality, sparse data and long computation time in traditional SVM classification algorithm based on term frequency-inverse document frequency (TF-IDF), we propose a novel hybrid system for Chinese text classification: CSVM, which is independent of the hand-designed features and domain knowledge. Firstly, the encoding words are done by constructing a text vocabulary of size m for the input language, and then quantise each word using 1-of-m encoding. Secondly, we exploit the convolutional neural network (CNN) to extract the morphological features of character vectors from each word, and then through large scale text material training the semantic feature of each word vectors are be obtained the semantic feature of each word vectors. Finally, the text classification is carried out with the SVM multiple classifier. The experimental results show that the CSVM algorithm is more effective than other traditional Chinese text classification algorithm.

English Else
Sentiment Classification for Chinese Reviews Based on Key Substring Features

Zhongwu Zhai,Hua Xu,Jun Li,Peifa Jia

DOI: https://doi.org/10.1109/nlpke.2009.5313782

2009-01-01

Abstract:One of the most widely-studied sub-problems of opinion mining is sentiment classification, which classifies evaluative texts as positive or negative to help people automatically identify the viewpoints underlying the online user-generated information. Most of the existing methods for sentiment classification ignore word sequence and unlabeled test documents' structural information. This paper proposes a transductive learning based algorithm considering both of these two types of information. The proposed algorithm is implemented by firstly selecting key substrings in the suffix tree constructed from the strings in training and unlabeled test documents and then converting each original text document to a bag of numbers of the key substrings. Finally, SVM is employed to classify the converted documents. Experiments on the open dataset (16,000 Chinese reviews) demonstrate promising performance of the proposed algorithm, the accuracy being over 93.15%, which is much better than the performance of the existing sentiment classification methods, such as n-gram features based classification algorithms. Experimental results also show that "tfidf-c" performs much better than other term weighting approaches in sentiment classification for large text corpus. In particular, the reasons behind the proposed algorithm's outstanding performance are further studied and analyzed in this paper. Moreover, the proposed algorithm can avoid the messy and rather artificial problem of defining word boundaries in Chinese language.
Combining Convolutional Neural Networks and Word Sentiment Sequence Features for Chinese Text Sentiment Classification

Zhao Chen,Ruifeng Xu,Lin Gui,Qin Lu

2015-01-01

Abstract:Combining Convolutional Neural Networks and Word Sentiment Sequence Features for Chinese Text Sentiment Classification Zhao Chen1, Ruifeng Xu1, Lin Gui1, Qin Lu2 (1. School of Computer Science and Technology, Harbin Institute of Technology Shenzhen Graduate School, Shenzhen, Guangdong, 518000, China; 2. Depart of Computing, The Hong Kong Polytechnic University, Hong Kong, China) Abstract: Recently, the classification approach based on word embedding and convolutional neural networks achieved good results in sentiment classification task. This approach is mainly based on the contextual features of the word embeddings without the consideration of the polarity of the words. Meanwhile, this approach lacks of the use of manually compiled sentiment lexicon resources. Target to these problems, this paper proposes a novel sentiment classification method which incorporates existing sentiment lexicon and convolution neural networks. In this word, the words in text are abstractly represented by using existing sentiment words. The convolutional neural networks are used to extract sequence features from the abstracted word embeddings. Finally, the sequence features are applied to sentiment classification. The evaluations on Chinese Opinion Analysis Evaluation 2014 dataset show that our proposed approach outperforms the convolutional neural networks model with word embedding features and Naïve Bayes Support Vector Machines.
Chinese Sentiment Analysis of MOOC Reviews Based on Word Vectors

Hua Yang

DOI: https://doi.org/10.1109/icaie53562.2021.00021

2021-06-01

Abstract:Sentiment analysis of MOOC reviews can reflect learners’ learning feelings, topics of interest and satisfaction with courses to platform managers and teachers, so as to help them make improvements and improve the quality of teaching. Potential learners can also find out the emotional information carried by the attributes of the courses they care about according to the reviews, which can be used as the basis for choosing courses or not. From the perspective of word vectors, this paper uses two word vector tools, Word2vec and Glove, to train MOOC review word vector data respectively, compares the impact of the two word vector tools on sentiment analysis, and visualizes the weight of attention. The experiment found that Word2vec was better than Glove on the whole, and the CBOW model under Word2vec had the best performance in binary classification, while the skip-gram model had the best performance in three classification. In addition, the effects of the two tools in finding similar words in new words was discussed. The experiment found that the Word2vec was better than the Glove, and the skip-gram model was more accurate than the CBOW model.
Modeling Blog User for Chinese Micro-Blog Sentiment Classification with Fuzzy SVM.

Xiao Sun,Chengcheng Li,Fuji Ren

DOI: https://doi.org/10.1109/ccis.2014.7175716

2014-01-01

Abstract:Micro-blog is easily posted and communicated because of its short text length. Meanwhile, its brevity limits the feature expression. This paper presents a blogger modeling method for Chinese micro-blog sentiment classification. The paper would adopt a multidimensional strategy to accomplish the feature extraction task. Compared with traditional feature extraction methods and combined with the interactive property of blog, the blog comment will be considered as the performance characteristic of micro-blog emotion expression. It comes to comments content and the relationship between commentators and micro-bloggers while it extracts the micro blogging text feature. At last, the paper adopts fuzzy SVM model to reduce the sensation of noise points and outliers on classification problem. The results demonstrate the effectiveness of the multidimensional feature extraction method on sentiment classification and the classify effect of fuzzy SVM is better than SVM.
A Parallel Two-Channel Emotion Classification Method for Chinese Text

Liu Na,Tao Cao,Shuchen Bai,Danqing Li

DOI: https://doi.org/10.1109/access.2024.3350190

IF: 3.9

2024-01-01

IEEE Access

Abstract:The complex structure of Chinese, including the dual-granularity feature of words and phrases, poses unique challenges for sentiment analysis, which is significantly different from the alphabetic word formation mechanism of English text. To capture Chinese semantics more accurately, an advanced parallel channel sentiment classification strategy is designed. In this study, the advanced pre-training model ERNIE with Word2vec is first adopted to enrich the embedding of Chinese text at word granularity. Following that, fine-grained features at the word level are extracted by a multi-window convolutional strategy, and the word vector sequences are deeply learnt using BiGRU network to ensure the comprehensive capture of contextual information in both directions. To further optimise the model, an Attention mechanism is introduced to ensure effective delivery of information and improve computational efficiency. After this series of innovative designs, the microblog comment dataset is used as an experimental case, and the hyperparameters are adjusted to determine the optimal parameters. Six comparison models are selected to verify the effectiveness of MsCBA on three datasets. The classification accuracies of the proposed model on the three datasets are 93.64%, 90.00% and 92.61%, respectively, which are better than the comparison models. This study provides an efficient and innovative approach for Chinese sentiment analysis, which sheds new light on the field of NLP.

computer science, information systems,telecommunications,engineering, electrical & electronic

Chinese comments sentiment classification based on word2vec and SVMperf.

Chinese Sentiment Classification Using A Neural Network Tool-Word2vec

Sentiment Classification for Chinese Reviews: a Comparison Between SVM and Semantic Approaches

Chinese Micro-Blog Emotion Classification By Exploiting Linguistic Features And Svmperf

A Multi-dimension and Multi-granularity Feature Fusion Method for Chinese Microblog Sentiment Classification

Sentiment Text Classification of Customers Reviews on the Web Based on SVM

Chinese Micro-Blog Emotion Classification by Exploiting Linguistic Features and SVM $$^{\Textit{perf}}$$

Combining Vector Space Features and Convolution Neural Network for Text Sentiment Analysis.

SVM-Based Sentiment Analysis Algorithm of Chinese Microblog under Complex Sentence Pattern

An Empirical Study on Sentiment Classification of Chinese Review using Word Embedding

Combining convolutional neural network and support vector machine for sentiment classification

Sentiment Classification for Chinese News Using Machine Learning Methods

Implementing Chinese new word discovery and POS tagging based on support vector machine

Sentiment Classification for Consumer Word-of-Mouth in Chinese: Comparison Between Supervised and Unsupervised Approaches

Feature representation based on sentimental orientation classification

Chinese text classification based on character-level CNN and SVM

Sentiment Classification for Chinese Reviews Based on Key Substring Features

Combining Convolutional Neural Networks and Word Sentiment Sequence Features for Chinese Text Sentiment Classification

Chinese Sentiment Analysis of MOOC Reviews Based on Word Vectors

Modeling Blog User for Chinese Micro-Blog Sentiment Classification with Fuzzy SVM.

A Parallel Two-Channel Emotion Classification Method for Chinese Text