Improving the Performance of Sentiment Analysis by Ensemble Hybrid Learning Algorithm With NLP And Cascaded Feature Extraction

Ayşe Berna Altınel Girgin; Sema Şahin

Makale detayları

Improving the Performance of Sentiment Analysis by Ensemble Hybrid Learning Algorithm With NLP And Cascaded Feature Extraction

Yazarlar:

Ayşe Berna Altınel Girgin | Sema Şahin

Yayıncı:

Marmara University

Ana Başlıklar:

Mühendislik

Dergi:

International Journal of Advances in Engineering and Pure Sciences Vol.35 Issue 1, pp.125-141

Belge Tipi:

Makale

Yayın Tarihi:

3/2023

Belge Dili:

İngilizce

Document ID (Org No.):

2001286649

pISSN:

2636-8277

Anahtar kelimeler:

Özet:

 Sentiment analysis is a challenging problem in Natural Language Processing since every language has its own character within several difficulties such as ambiguity, synonymy, negative suffixesetc. Since words with ambiguity can have different sentiment scores depending on the meaning they have in their corresponding context, we accomplished a study on Turkish language to determine whether the polarity scores of these polysemous words may differ according to their meaning. For a word with ambiguity, we first made a polarity calculation module to calculate its polarity score. This way, we calculated the polarity scores of 100 Turkish polysemous words. Then, since negation directly affects the correct meaning of the word in the sentiment analysis, a negation handler module is also implemented. After that, we prepared a sentiment polarity corpus which consists of 159,876 Turkish words including 100 Turkish polysemous words. Actually, the main purpose of this study is to detect sentiment polarity of Turkish texts by considering and building a specialized module for polysemous words. In short, we built a system for Turkish sentiment polarity detection task including these modules: 1 Pre-processing, 2 Polarity Calculation Module, 3 Negation Handling Module, 4 Feature Generation Module, and 5 Classification Module. According to our knowledge, this is the first study which includes all of these modules in one Turkish sentiment analysis task. Finally, we conducted this corpus using an ensemble hybrid regularized learning algorithm on two self-collected Twitter-datasets. Experimental results show that the suggested approach improves the classification performance on Turkish sentiment analysis task.

 Sentiment analysis is a challenging problem in Natural Language Processing since every language has its own character within several difficulties such as ambiguity, synonymy, negative suffixesetc. Since words with ambiguity can have different sentiment scores depending on the meaning they have in their corresponding context, we accomplished a study on Turkish language to determine whether the polarity scores of these polysemous words may differ according to their meaning. For a word with ambiguity, we first made a polarity calculation module to calculate its polarity score. This way, we calculated the polarity scores of 100 Turkish polysemous words. Then, since negation directly affects the correct meaning of the word in the sentiment analysis, a negation handler module is also implemented. After that, we prepared a sentiment polarity corpus which consists of 159,876 Turkish words including 100 Turkish polysemous words. Actually, the main purpose of this study is to detect sentiment polarity of Turkish texts by considering and building a specialized module for polysemous words. In short, we built a system for Turkish sentiment polarity detection task including these modules: 1 Pre-processing, 2 Polarity Calculation Module, 3 Negation Handling Module, 4 Feature Generation Module, and 5 Classification Module. According to our knowledge, this is the first study which includes all of these modules in one Turkish sentiment analysis task. Finally, we conducted this corpus using an ensemble hybrid regularized learning algorithm on two self-collected Twitter-datasets. Experimental results show that the suggested approach improves the classification performance on Turkish sentiment analysis task.

 Sentiment analysis is a challenging problem in Natural Language Processing since every language has its own character within several difficulties such as ambiguity, synonymy, negative suffixesetc. Since words with ambiguity can have different sentiment scores depending on the meaning they have in their corresponding context, we accomplished a study on Turkish language to determine whether the polarity scores of these polysemous words may differ according to their meaning. For a word with ambiguity, we first made a polarity calculation module to calculate its polarity score. This way, we calculated the polarity scores of 100 Turkish polysemous words. Then, since negation directly affects the correct meaning of the word in the sentiment analysis, a negation handler module is also implemented. After that, we prepared a sentiment polarity corpus which consists of 159,876 Turkish words including 100 Turkish polysemous words. Actually, the main purpose of this study is to detect sentiment polarity of Turkish texts by considering and building a specialized module for polysemous words. In short, we built a system for Turkish sentiment polarity detection task including these modules: 1 Pre-processing, 2 Polarity Calculation Module, 3 Negation Handling Module, 4 Feature Generation Module, and 5 Classification Module. According to our knowledge, this is the first study which includes all of these modules in one Turkish sentiment analysis task. Finally, we conducted this corpus using an ensemble hybrid regularized learning algorithm on two self-collected Twitter-datasets. Experimental results show that the suggested approach improves the classification performance on Turkish sentiment analysis task.

 Her doğal dilin belirsizlik, eşanlamlılık, olumsuz ekler vb. gibi çeşitli zorluklar içeren bir karakterinin olmasından ötürü Duygu Analizi, Doğal Dil İşlemede zorlu bir problemdir. Belirsizlik içeren eşsesli sözcükler, bulundukları bağlamdaki anlamlarına göre farklı duygu polaritesi puanlarına sahip olabildiği için, bu eşsesli sözcüklerin anlamlarına göre farklı duygu polaritesi puanları aldıklarını analiz eden Türkçe üzerine bir çalışma gerçekleştirdik. Belirsizliği olan bir kelime için, polarite puanını hesaplamak için önce bir polarite hesaplama modülü yaptık. Bu şekilde 100 Türkçe eşsesli kelimenin polarite puanlarını hesapladık. Ardından, duygu analizinde olumsuzlama kelimenin doğru anlamını doğrudan etkilediğinden, olumsuzluk işleyici modülü de gerçekleştirdik. Daha sonra 100 Türkçe eşsesli kelime olmak üzere 159.876 Türkçe kelimeden oluşan duygu polaritesi sözlüğü hazırladık. Aslında bu çalışmanın temel amacı, eşsesli sözcükler için özel bir modül oluşturarak Türkçe metinlerdeki duygu polaritesini tespit etmektir. Kısaca, Türkçe duygu polarite tespiti görevi için şu modülleri içeren bir sistem kurduk: 1 Ön işleme, 2 Polarite Hesaplama Modülü, 3 Olumsuzlama Modülü, 4 Özellik Oluşturma Modülü ve 5 Sınıflandırma Modülü. Bildiğimiz kadarıyla bu çalışma, Türkçe duygu analizi görevi için tüm bu modülleri bir arada içeren ilk çalışmadır. Son olarak, bu sözlüğü, kendi topladığımız iki Twitter veri kümesinde bir topluluk hibrit öğrenme algoritması kullanarak analiz ettik. Deneysel sonuçlar, önerilen yaklaşımın Türkçe duygu analizi görevinde sınıflandırma performansını arttırdığını göstermektedir.

Makale detayları

Improving the Performance of Sentiment Analysis by Ensemble Hybrid Learning Algorithm With NLP And Cascaded Feature Extraction

Yazarlar:

Ayşe Berna Altınel Girgin | Sema Şahin

Yayıncı:

Marmara University

Ana Başlıklar:

Mühendislik

Dergi:

International Journal of Advances in Engineering and Pure Sciences Vol.35 Issue 1, pp.125-141

Belge Tipi:

Makale

Yayın Tarihi:

3/2023

Belge Dili:

İngilizce

Document ID (Org No.):

2001286649

pISSN:

2636-8277

Sentiment analysis

word ambiguity

machine learning

hybrid learning algorithm

LSTM

Duygu analizi

kelime anlam bulanıklığı

makine öğrenmesi

hibrit öğrenme algoritması

LSTM

Sayı

Makale detayları

Yazarlar: Ayşe Berna Altınel Girgin | Sema Şahin

Yayıncı: Marmara University

Ana Başlıklar: Mühendislik

Dergi: International Journal of Advances in Engineering and Pure Sciences Vol.35 Issue 1, pp.125-141

Belge Tipi: Makale

Yayın Tarihi: 3/2023

Belge Dili: İngilizce

Document ID (Org No.): 2001286649

pISSN: 2636-8277

Sayı

Yazarlar:

Ayşe Berna Altınel Girgin | Sema Şahin

Yayıncı:

Marmara University

Ana Başlıklar:

Mühendislik

Dergi:

International Journal of Advances in Engineering and Pure Sciences Vol.35 Issue 1, pp.125-141

Belge Tipi:

Makale

Yayın Tarihi:

3/2023

Belge Dili:

İngilizce

Document ID (Org No.):

2001286649

pISSN:

2636-8277