Sentiment Analysis in Turkish Texts: The Effect of Deep Learning Approaches and Preprocessing Processes on Model Performance

Yasin Görmez; Halil Arslan; Bilal Atak

Makale detayları

Sentiment Analysis in Turkish Texts: The Effect of Deep Learning Approaches and Preprocessing Processes on Model Performance

Yazarlar:

Yasin Görmez | Halil Arslan | Bilal Atak

Yayıncı:

Fırat University

Ana Başlıklar:

Derin Öğrenme | Doğal Dil İşleme

Dergi:

Fırat Üniversitesi Mühendislik Bilimleri Dergisi Vol.36 Issue 1, pp.509-520

Belge Tipi:

Makale

Yayın Tarihi:

3/2024

Belge Dili:

Türkçe

Document ID (Org No.):

2001302461

pISSN:

1308-9072

Anahtar kelimeler:

Özet:

 Nowadays, with the increased use of computers, a surge in data production has emerged, making data access more convenient. In this context, a substantial amount of textual data is generated on e-commerce sites, social media, and various electronic platforms. Analyzing and extracting meaningful insights from this amassed data proves valuable for numerous institutions, organizations, and individuals. Sentiment analysis is a commonly employed technique to derive sentiments from textual data, and contemporary sentiment analysis models often leverage the high performance offered by deep learning approaches. Prior to model training, several pre-processing steps are typically applied to the text data. In this study, three distinct deep learning approaches were proposed for sentiment analysis. These models were analyzed on two different datasets: winvoker and Beyazperde. Hyper-parameters and depth of models were optimized using the Bayesian optimization method to enhance the accuracy of model. Additionally, the impact of various pre-processing techniques on model performance were assessed. When non-preprocessed data is utilized, the models trained on the winvoker dataset achieve an accuracy of 94.16, while those trained on the Beyazperde dataset reach 86.64. With the application of pre-processing, these accuracies improve to 94.64 for the winvoker dataset and 89.08 for the Beyazperde dataset. Based on these findings, it was concluded that the effect of pre-processing decreased and the accuracy was higher for the winvoker data set with a higher number of samples.

 Nowadays, with the increased use of computers, a surge in data production has emerged, making data access more convenient. In this context, a substantial amount of textual data is generated on e-commerce sites, social media, and various electronic platforms. Analyzing and extracting meaningful insights from this amassed data proves valuable for numerous institutions, organizations, and individuals. Sentiment analysis is a commonly employed technique to derive sentiments from textual data, and contemporary sentiment analysis models often leverage the high performance offered by deep learning approaches. Prior to model training, several pre-processing steps are typically applied to the text data. In this study, three distinct deep learning approaches were proposed for sentiment analysis. These models were analyzed on two different datasets: winvoker and Beyazperde. Hyper-parameters and depth of models were optimized using the Bayesian optimization method to enhance the accuracy of model. Additionally, the impact of various pre-processing techniques on model performance were assessed. When non-preprocessed data is utilized, the models trained on the winvoker dataset achieve an accuracy of 94.16, while those trained on the Beyazperde dataset reach 86.64. With the application of pre-processing, these accuracies improve to 94.64 for the winvoker dataset and 89.08 for the Beyazperde dataset. Based on these findings, it was concluded that the effect of pre-processing decreased and the accuracy was higher for the winvoker data set with a higher number of samples.

 Nowadays, with the increased use of computers, a surge in data production has emerged, making data access more convenient. In this context, a substantial amount of textual data is generated on e-commerce sites, social media, and various electronic platforms. Analyzing and extracting meaningful insights from this amassed data proves valuable for numerous institutions, organizations, and individuals. Sentiment analysis is a commonly employed technique to derive sentiments from textual data, and contemporary sentiment analysis models often leverage the high performance offered by deep learning approaches. Prior to model training, several pre-processing steps are typically applied to the text data. In this study, three distinct deep learning approaches were proposed for sentiment analysis. These models were analyzed on two different datasets: winvoker and Beyazperde. Hyper-parameters and depth of models were optimized using the Bayesian optimization method to enhance the accuracy of model. Additionally, the impact of various pre-processing techniques on model performance were assessed. When non-preprocessed data is utilized, the models trained on the winvoker dataset achieve an accuracy of 94.16, while those trained on the Beyazperde dataset reach 86.64. With the application of pre-processing, these accuracies improve to 94.64 for the winvoker dataset and 89.08 for the Beyazperde dataset. Based on these findings, it was concluded that the effect of pre-processing decreased and the accuracy was higher for the winvoker data set with a higher number of samples.

 Günümüzde bilgisayar kullanımın artması ile birlikte insanlar daha fazla veri üretmeye başlamış ve verilere ulaşım kolaylaşmıştır. Bu bağlamda e-ticaret sitelerinde, sosyal medyada ya da diğer elektronik platformlarda çok fazla metin verisi üretilmiştir. Toplanan bu verilerin analiz edilerek anlamlandırılması birçok kurum, kuruluş ya da birey için faydalı bilgiler sağlamaktadır. Bu amaç doğrultusunda duygu analizi günümüzde sıklıkla uygulanmaktadır. Duygu analizi modellerinde derin öğrenme yaklaşımları oldukça yüksek performans göstermekte ve model eğitimi yapılmadan önce metinlere birkaç ön işlem uygulanmaktadır. Bu çalışmada duygu analizi için, evrişimsel sinir ağı, Transfomer ve hibrit olmak üzere üç farklı derin öğrenme yaklaşımı önerilmiş ve modeller winvoker ve Beyazperde olmak üzere iki farklı veri seti kullanılarak analiz edilmiştir. Modellerin doğruluğunu artırmak için hiper-parametreleri ve model derinliklileri Bayesian optimizasyon yöntemi kullanılarak optimize edilmiştir. Ön işlem süreçlerinin model performansına etkisini ölçmek için veri setlerine çeşitli ön işlem yapılarak analizler tekrar edilmiştir. Ön işlem uygulanmamış veriler kullanıldığında, winvoker veri seti ile eğitilen modellerde 94,16, Beyazperde veri seti ile eğitilen modellerde ise 86,64 doğruluğa ulaşılmıştır. Ön işlem uygulandığında ise bu başarı oranları, winvoker veri seti ile eğitilen modellerde 94,64, Beyazperde veri seti ile eğitilen modellerde ise 89,08 değerlerine ulaşmıştır. Bu sonuçlar doğrultusunda örnek sayısı daha fazla olan winvoker veri seti için ön işlemlerin etkisinin azaldığı ve doğruluğun daha yüksek olduğu sonucu çıkarılmıştır.

Makale detayları

Sentiment Analysis in Turkish Texts: The Effect of Deep Learning Approaches and Preprocessing Processes on Model Performance

Yazarlar:

Yasin Görmez | Halil Arslan | Bilal Atak

Yayıncı:

Fırat University

Ana Başlıklar:

Derin Öğrenme | Doğal Dil İşleme

Dergi:

Fırat Üniversitesi Mühendislik Bilimleri Dergisi Vol.36 Issue 1, pp.509-520

Belge Tipi:

Makale

Yayın Tarihi:

3/2024

Belge Dili:

Türkçe

Document ID (Org No.):

2001302461

pISSN:

1308-9072

sentiment analysis

deep learning

text pre-processing

convolutional neural networks

transformer networksduygu analizi

derin öğrenme

metin ön işlem

evrişimsel sinir ağları

transformer ağları

Sayı

Makale detayları

Yazarlar: Yasin Görmez | Halil Arslan | Bilal Atak

Yayıncı: Fırat University

Ana Başlıklar: Derin Öğrenme | Doğal Dil İşleme

Dergi: Fırat Üniversitesi Mühendislik Bilimleri Dergisi Vol.36 Issue 1, pp.509-520

Belge Tipi: Makale

Yayın Tarihi: 3/2024

Belge Dili: Türkçe

Document ID (Org No.): 2001302461

pISSN: 1308-9072

Sayı

Yazarlar:

Yasin Görmez | Halil Arslan | Bilal Atak

Yayıncı:

Fırat University

Ana Başlıklar:

Derin Öğrenme | Doğal Dil İşleme

Dergi:

Fırat Üniversitesi Mühendislik Bilimleri Dergisi Vol.36 Issue 1, pp.509-520

Belge Tipi:

Makale

Yayın Tarihi:

3/2024

Belge Dili:

Türkçe

Document ID (Org No.):

2001302461

pISSN:

1308-9072