Opinion mining in Persian language using a hybrid feature extraction approach based on convolutional neural network

Zobeidi, Shima; Naderan, Marjan; Alavi, Seyyed Enayatallah

doi:10.1007/s11042-019-07993-4

Cited by 17 publications

(7 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In both models, the vector related to each word is determined based on the word's meaning and the relationship between the word and the surroundings. 27 CBOW is used in this research to determine the vector related to each word. After determining the vector related to each word, the sentences' matrix is created based on the word sequence.…”

Section: Proposed Methods Preprocessingmentioning

confidence: 99%

See 1 more Smart Citation

Deep aspect extraction and classification for opinion mining in e‐commerce applications using convolutional neural network feature extraction followed by long short term memory attention model

Sharbatian,

Moattar

2023

Applied AI Letters

View full text Add to dashboard Cite

Users of e‐commerce websites review different aspects of a product in the comment section. In this research, an approach is proposed for opinion aspect extraction and recognition in selling systems. We have used the users' opinions from the Digikala website (www.Digikala.com), which is an Iranian e‐commerce company. In this research, a language‐independent framework is proposed that is adjustable to other languages. In this regard, after necessary text processing and preparation steps, the existence of an aspect in an opinion is determined using deep learning algorithms. The proposed model combines Convolutional Neural Network (CNN) and long‐short‐term memory (LSTM) deep learning approaches. CNN is one of the best algorithms for extracting latent features from data. On the other hand, LSTM can detect latent temporal relationships among different words in a text due to its memory ability and attention model. The approach is evaluated on six classes of opinion aspects. Based on the experiments, the proposed model's accuracy, precision, and recall are 70%, 60%, and 85%, respectively. The proposed model was compared in terms of the above criteria with CNN, Naive Bayes, and SVM algorithms and showed satisfying performance.

show abstract

Section: Proposed Methods Preprocessingmentioning

confidence: 99%

“…The significant note in this research is using POS related to each word. The authors in Zobeidi et al 27 gather the users' comments about mobile phones and digital cameras. They use CNN to extract the features and CBOW to determine the text opinions as a numerical matrix.…”

Section: Related Workmentioning

confidence: 99%

Deep aspect extraction and classification for opinion mining in e‐commerce applications using convolutional neural network feature extraction followed by long short term memory attention model

Sharbatian,

Moattar

2023

Applied AI Letters

View full text Add to dashboard Cite

show abstract

“…Bokaee Nezhad and Deihimi [26] examined various models for Persian sentiment analysis, including CNN, CNN-LSTM and GRU, and showed that CNN-LSTM performs the best based on precision, recall and f-score metrics. A similar approach which includes the combination of CNN, LSTM and BiLSTM networks has also been proposed by Zobeidi et al [27] for fine-grained Persian sentiment analysis. Basiri and Kabiri [28] proposed a new hybrid system for opinion mining in the Persian language.…”

Section: Related Workmentioning

confidence: 99%

Deep Persian sentiment analysis: Cross-lingual training for low-resource languages

Ghasemi

Asli

Momtazi

2020

Journal of Information Science

View full text Add to dashboard Cite

With the advent of deep neural models in natural language processing tasks, having a large amount of training data plays an essential role in achieving accurate models. Creating valid training data, however, is a challenging issue in many low-resource languages. This problem results in a significant difference between the accuracy of available natural language processing tools for low-resource languages compared with rich languages. To address this problem in the sentiment analysis task in the Persian language, we propose a cross-lingual deep learning framework to benefit from available training data of English. We deployed cross-lingual embedding to model sentiment analysis as a transfer learning model which transfers a model from a rich-resource language to low-resource ones. Our model is flexible to use any cross-lingual word embedding model and any deep architecture for text classification. Our experiments on English Amazon dataset and Persian Digikala dataset using two different embedding models and four different classification networks show the superiority of the proposed model compared with the state-of-the-art monolingual techniques. Based on our experiment, the performance of Persian sentiment analysis improves 22% in static embedding and 9% in dynamic embedding. Our proposed model is general and language-independent; that is, it can be used for any low-resource language, once a cross-lingual embedding is available for the source–target language pair. Moreover, by benefitting from word-aligned cross-lingual embedding, the only required data for a reliable cross-lingual embedding is a bilingual dictionary that is available between almost all languages and the English language, as a potential source language.

show abstract

“…The confusion matrices are applied to validate the classifier's performance and accuracy measures [54], [61]. In this evaluation, Precision, Recall, F-Measure, and Accuracy matrix are adopted with binary classes of positive and negative sentiments.…”

Section: ) Performance Evaluation With Confusion Matricesmentioning

confidence: 99%

GAWA–A Feature Selection Method for Hybrid Sentiment Classification

et al. 2020

View full text Add to dashboard Cite

Sentiment analysis or opinion mining is the key to natural language processing for the extraction of useful information from the text documents of numerous sources. Several different techniques, i.e., simple rule-based to lexicon-based and more sophisticated machine learning algorithms, have been widely used with different classifiers to get the factual analysis of sentiment. However, lexicon-based sentiment classification is still suffering from low accuracies, mainly due to the deficiency of domain-oriented competitive dictionaries. Similarly, machine learning-based sentiment is also tackling the accuracy constraints because of feature ambiguity from social data. One of the best ways to deal with the accuracy issue is to select the best feature-set and reduce the volume of the feature. This paper proposes a method (namely, GAWA) for feature selection by utilizing the Wrapper Approaches (WA) to select the premier features and the Genetic Algorithm (GA) to reduce the size of the premier features. The novelty of this work is the modified fitness function of heuristic GA to compute the optimal features by reducing the redundancy for better accuracy. This work aims to present a comprehensive model of hybrid sentiment by using the proposed method, GAWA. It will be valued in developing a new approach for the selection of feature-set with a better accuracy level. The experiments revealed that these techniques could reduce the feature-set upto 61.95% without negotiating the accuracy level. The new optimal feature sets enhanced the efficiency of the Naïve Bayes algorithm up to 92%. This work is compared with the conventional method of feature selection and concluded the 11% better accuracy than PCA and 8% better than PSO. Furthermore, the results are compared with the literature work and found that the proposed method outperformed the previous research.

show abstract

Opinion mining in Persian language using a hybrid feature extraction approach based on convolutional neural network

Cited by 17 publications

References 36 publications

Deep aspect extraction and classification for opinion mining in e‐commerce applications using convolutional neural network feature extraction followed by long short term memory attention model

Deep aspect extraction and classification for opinion mining in e‐commerce applications using convolutional neural network feature extraction followed by long short term memory attention model

Deep Persian sentiment analysis: Cross-lingual training for low-resource languages

GAWA–A Feature Selection Method for Hybrid Sentiment Classification

Contact Info

Product

Resources

About