Detection of Aspect Category and Aspect-Based Sentiment Analysis on Restaurant Customers' Feedbacks

Murat Fatih Tuna; Mesut Polatgil; Oğuz Kaynar

Makale detayları

Detection of Aspect Category and Aspect-Based Sentiment Analysis on Restaurant Customers' Feedbacks

Yazarlar:

Murat Fatih Tuna | Mesut Polatgil | Oğuz Kaynar

Yayıncı:

Süleyman Demirel University

Ana Başlıklar:

İşletme

Dergi:

Süleyman Demirel Üniversitesi Vizyoner Dergisi Vol.14 Issue 40, pp.1205-1221

Belge Tipi:

Makale

Yayın Tarihi:

11/2023

Belge Dili:

Türkçe

Document ID (Org No.):

2001303922

pISSN:

1308-9552

Anahtar kelimeler:

Özet:

 Today, there are many channels where consumers can share their ideas about products and services. These opinions are usually in text format due to the nature of the feedback. Sentiment analysis is a topic that has gained importance in recent years, especially in text-based information sources. Aspect-based Sentiment Analysis, which is a more sensitive sentiment analysis technique, is the task of determining the aspect term, aspect category and sentiment class in a sentence. A data set consisting of the comments of restaurant customers presented to the competitors in the Semeval ABSA competition is used in the study. Using Word2vec, Glove, Fastext and Bert methods, the aspect term, aspect category and sentiment class are determined on the data set. The hypothesis is tested whether combining the word vector and the sentence vector can improve classification success for ABSA. In the classification made with four different vector methods, Fasttext method with 0.78 F1 score for the target term, Fasttext with 0.57 F1 score for the target category, and Bert method with 0.76 F1 score for the sentiment class have the most successful results. These results are compared with studies in the literature for different data sets and different languages. As a result, it is determined that Fasttext and Bert representation methods give successful results in sentiment analysis of target-based Turkish language texts.

 Today, there are many channels where consumers can share their ideas about products and services. These opinions are usually in text format due to the nature of the feedback. Sentiment analysis is a topic that has gained importance in recent years, especially in text-based information sources. Aspect-based Sentiment Analysis, which is a more sensitive sentiment analysis technique, is the task of determining the aspect term, aspect category and sentiment class in a sentence. A data set consisting of the comments of restaurant customers presented to the competitors in the Semeval ABSA competition is used in the study. Using Word2vec, Glove, Fastext and Bert methods, the aspect term, aspect category and sentiment class are determined on the data set. The hypothesis is tested whether combining the word vector and the sentence vector can improve classification success for ABSA. In the classification made with four different vector methods, Fasttext method with 0.78 F1 score for the target term, Fasttext with 0.57 F1 score for the target category, and Bert method with 0.76 F1 score for the sentiment class have the most successful results. These results are compared with studies in the literature for different data sets and different languages. As a result, it is determined that Fasttext and Bert representation methods give successful results in sentiment analysis of target-based Turkish language texts.

 Today, there are many channels where consumers can share their ideas about products and services. These opinions are usually in text format due to the nature of the feedback. Sentiment analysis is a topic that has gained importance in recent years, especially in text-based information sources. Aspect-based Sentiment Analysis, which is a more sensitive sentiment analysis technique, is the task of determining the aspect term, aspect category and sentiment class in a sentence. A data set consisting of the comments of restaurant customers presented to the competitors in the Semeval ABSA competition is used in the study. Using Word2vec, Glove, Fastext and Bert methods, the aspect term, aspect category and sentiment class are determined on the data set. The hypothesis is tested whether combining the word vector and the sentence vector can improve classification success for ABSA. In the classification made with four different vector methods, Fasttext method with 0.78 F1 score for the target term, Fasttext with 0.57 F1 score for the target category, and Bert method with 0.76 F1 score for the sentiment class have the most successful results. These results are compared with studies in the literature for different data sets and different languages. As a result, it is determined that Fasttext and Bert representation methods give successful results in sentiment analysis of target-based Turkish language texts.

 Günümüzde tüketicilerin ürün ve hizmetler konusunda fikir paylaşabilecekleri birçok mecra bulunmaktadır. Bu fikirler, geri bildirimin yapısı itibariyle genellikle metin formatındadır. Duygu analizi, metin tabanlı bilgi kaynaklarında son yıllarda önem kazanan bir konudur. Daha hassas bir duygu analiz türü olan Hedef Tabanlı Duygu Analizi bir cümle içerisinde hedef terim, hedef kategori ve duygu sınıfının belirlenmesi işidir. Bu çalışmada Semeval ABSA yarışmasında yarışmacılara sunulan restoran müşterilerine ait yorumlardan oluşan bir veri seti kullanılmıştır. Word2vec, Glove, Fastext ve Bert yöntemleri kullanılarak veri seti üzerinde hedef terim, hedef kategori ve duygu sınıfının belirlenmesi işlemi gerçekleştirilmiştir. Kelimeyi vektörü ile cümle vektörünün birleştirilmesi ABSA için sınıflandırma başarısını artırıp artıramayacağı hipotezi test edilmiştir. Dört farklı vektör yöntemi ile yapılan sınıflandırmada hedef terim için 0,78 F1 skoru ile Fasttext yöntemi, hedef kategori için 0,57 F1 skoru ile Fasttext ve duygu sınıfı için 0,76 F1 skoru ile Bert yöntemi en başarılı sonuçları vermiştir. Bu sonuçlar literatürde farklı veri setleri ve farklı diller için yapılan çalışmalarla kıyaslanmıştır. Sonuç olarak Fasttext ve Bert temsil yöntemlerinin hedef tabanlı Türkçe dilindeki metinlerin duygu analizinde başarılı sonuçlar verdiği tespit edilmiştir.

Makale detayları

Detection of Aspect Category and Aspect-Based Sentiment Analysis on Restaurant Customers' Feedbacks

Yazarlar:

Murat Fatih Tuna | Mesut Polatgil | Oğuz Kaynar

Yayıncı:

Süleyman Demirel University

Ana Başlıklar:

İşletme

Dergi:

Süleyman Demirel Üniversitesi Vizyoner Dergisi Vol.14 Issue 40, pp.1205-1221

Belge Tipi:

Makale

Yayın Tarihi:

11/2023

Belge Dili:

Türkçe

Document ID (Org No.):

2001303922

pISSN:

1308-9552

Restaurant Reviews

Aspect Based Sentiment Analysis

Fasttext

BertRestoran yorumları

Hedef Tabanlı Duygu Analizi

Fasttext

Bert

Sayı

Makale detayları

Yazarlar: Murat Fatih Tuna | Mesut Polatgil | Oğuz Kaynar

Yayıncı: Süleyman Demirel University

Ana Başlıklar: İşletme

Dergi: Süleyman Demirel Üniversitesi Vizyoner Dergisi Vol.14 Issue 40, pp.1205-1221

Belge Tipi: Makale

Yayın Tarihi: 11/2023

Belge Dili: Türkçe

Document ID (Org No.): 2001303922

pISSN: 1308-9552

Sayı

Yazarlar:

Murat Fatih Tuna | Mesut Polatgil | Oğuz Kaynar

Yayıncı:

Süleyman Demirel University

Ana Başlıklar:

İşletme

Dergi:

Süleyman Demirel Üniversitesi Vizyoner Dergisi Vol.14 Issue 40, pp.1205-1221

Belge Tipi:

Makale

Yayın Tarihi:

11/2023

Belge Dili:

Türkçe

Document ID (Org No.):

2001303922

pISSN:

1308-9552