Sentiment Analysis in Turkish with Deep Learning

Demirci, Gozde Merve; Keskin, Şeref Recep; Doğan, Gülüstan

doi:10.1109/bigdata47090.2019.9006066

Cited by 22 publications

(14 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This model, which we refer to it as Turkish Deep, contains 6 dense layers, where each one has different output dimensions and a 0.5 Dropout is allocated after the second and the fourth layer, while 0.2 Dropout is allocated after the third one. The performance of multiple preprocessing steps and feature represents were investigated also in [12]. Overall results showed that the deep learning method has the potential to build a better solution for sentiment analysis.…”

Section: Cms Recent Developments and Studiesmentioning

confidence: 99%

“…It is worth mentioning that the proposed approach was able to achieve on average 5% as improvement comparing to the RF, which was obtained as a result of integrating both the developed preprocessing model and the Word2Vec. In this experiment, the performance of the proposed approach was compared with the approaches presented in [12] and [25]. In detail, we have re-implemented the Empirical approach [25], which uses the TF-IDF to represent the text features.…”

Section: Experiments 3: Performance Of the Proposed Systemmentioning

confidence: 99%

“…Also, we have re-implemented the CNN model presented in [12], which was built for processing the Turkish sentiment analysis task. As mentioned in the related work section, this model, which we refer to it as "Turkish Deep", contains 6 dense layers, where each one has different output dimensions and a 0.5 Dropout is allocated after the second and the fourth layer, while 0.2 Dropout is allocated after the third one.…”

Section: Experiments 3: Performance Of the Proposed Systemmentioning

confidence: 99%

“…Related to the CNN model of [12], it has achieved the second-best for the first and third datasets, however, as mentioned in the related work section, CNN models are currently the state-of-the-art models for many classification tasks including the text classification. We believe that CNN achieved this results due to lack of a large number of training samples, wherein the case of the used datasets the generated features during the training of CNN are not generalizable enough mainly due to the large number of the CNN's variables(weights) that need to be trained.…”

Section: Experiments 3: Performance Of the Proposed Systemmentioning

confidence: 99%

“…translation), classification and many other tasks, using such technique requires the availability of a large number of training data which is not available yet, i.e, our constructed datasets is the first one related to disasters for the Turkish language. Overall, unlike the approach of [12], which uses Convolutional Neural Networks, our approach does not require a huge amount of training data to achieve an acceptable performance.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Efficient Turkish Text Classification Approach for Crisis Management Systems

Alqaraleh

2021

Gazi University Journal of Science

View full text Add to dashboard Cite

show abstract

Section: Cms Recent Developments and Studiesmentioning

confidence: 99%

Section: Experiments 3: Performance Of the Proposed Systemmentioning

confidence: 99%

Section: Experiments 3: Performance Of the Proposed Systemmentioning

confidence: 99%

Section: Experiments 3: Performance Of the Proposed Systemmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Efficient Turkish Text Classification Approach for Crisis Management Systems

Alqaraleh

2021

Gazi University Journal of Science

View full text Add to dashboard Cite

show abstract

A novel COVID‐19 sentiment analysis in Turkish based on the combination of convolutional neural network and bidirectional long‐short term memory on Twitter

Kabakuş

2022

Concurrency and Computation

View full text Add to dashboard Cite

The whole world has been experiencing the COVID‐19 pandemic since December 2019. During the pandemic, a new life has been started by necessity where people have extensively used social media to express their feelings, and find information. Twitter was used as the source of what people have shared regarding the COVID‐19 pandemic. Sentiment analysis deals with the extraction of the sentiment of a given text. Most of the related works deal with sentiment analysis in English, while studies for Turkish sentiment analysis lack in the research field. To this end, a novel sentiment analysis model based on the combination of convolutional neural network and bidirectional long short‐term memory was proposed in this study. The proposed deep neural network model was trained on the constructed Twitter dataset, which consists of 15k Turkish tweets regarding the COVID‐19 pandemic, to classify a given tweet into three sentiment classes, namely, (i) positive, (ii) negative, and (iii) neutral. A set of experiments were conducted for the evaluation of the proposed model. According to the experimental result, the proposed model obtained an accuracy as high as 97.895%, which outperformed the state‐of‐the‐art baseline models for sentiment analysis of tweets in Turkish.

show abstract

Sentiment Analysis Using State of the Art Machine Learning Techniques

Balci

Demirci

Demirhan

et al. 2022

Digital Interaction and Machine Intelligence

Self Cite

View full text Add to dashboard Cite

Sentiment analysis is one of the essential and challenging tasks in the Artificial Intelligence field due to the complexity of the languages. Models that use rule-based and machine learning-based techniques have become popular. However, existing models have been under-performing in classifying irony, sarcasm, and subjectivity in the text. In this paper, we aim to deploy and evaluate the performances of the State-of-the-Art machine learning sentiment analysis techniques on a public IMDB dataset. The dataset includes many samples of irony and sarcasm. Long-short term memory (LSTM), bag of tricks (BoT), convolutional neural networks (CNN), and transformer-based models are developed and evaluated. In addition, we have examined the effect of hyper-parameters on the accuracy of the models.

show abstract

Sentiment Analysis in Turkish with Deep Learning

Cited by 22 publications

References 13 publications

Efficient Turkish Text Classification Approach for Crisis Management Systems

Efficient Turkish Text Classification Approach for Crisis Management Systems

A novel COVID‐19 sentiment analysis in Turkish based on the combination of convolutional neural network and bidirectional long‐short term memory on Twitter

Sentiment Analysis Using State of the Art Machine Learning Techniques

Contact Info

Product

Resources

About