Towards Accurate Deceptive Opinions Detection Based on Word Order-Preserving CNN

Abstract

Nowadays, deep learning has been widely used. In natural language learning, the analysis of complex semantics has been achieved because of its high degree of flexibility. The deceptive opinions detection is an important application area in deep learning model, and related mechanisms have been given attention and researched. On-line opinions are quite short, varied types and content. In order to effectively identify deceptive opinions, we need to comprehensively study the characteristics of deceptive opinions, and explore novel characteristics besides the textual semantics and emotional polarity that have been widely used in text analysis. The detection mechanism based on deep learning has better self-adaptability and can effectively identify all kinds of deceptive opinions. In this paper, we optimize the convolution neural network model by embedding the word order characteristics in its convolution layer and pooling layer, which makes convolution neural network more suitable for various text classification and deceptive opinions detection. The TensorFlow-based experiments demonstrate that the detection mechanism proposed in this paper achieve more accurate deceptive opinion detection results.

Fake Reviews Detection: A Survey

“…Furthermore, in mixed-domain, CNN performed better than LSTM. Motivated by this, Zhao, et al [16] introduced a word order-preserving CNN method for detecting fake reviews. They used word 2vec and the word order reserving pooling method rather than the original max pooling to generate a word vector.…”

“…In contrast, CNN was a more efficient model for classifying short text reviews. CNN had a shorter training time, while RNN was more efficient for long texts [16]. However, the handannotated technique requires a lot of manpower.…”

Sentiment analysis using deep learning approaches: an overview

“…CNN based models. Zhao et al [179] suggested a CNN model for deceptive opinion spam detection. The proposed model extended the CNN with word order preserving pooling layer, which allows to keep the order of words in a sentence while analyzing deceptive opinion spam.…”

Text Mining and Data Dimesion Reduction Approach on Consumer Comments

“…Bu sayede sinema filmi veya benzeri ürünler için yapılan yorumların polarizasyonu sağlanmıştır. Her iki yaklaşımın kıyaslaması da ek olarak yapılmıştır (Singh, 2014 (Zhao, 2018).…”

“…İlgili çalışmaların anlatıldığı kısımda da ifade edildiği üzere, Zubrinic'in 2018'de yayınladığı tüketici Duygu analizi çalışmasının, ikili (binary) bir veride elde ettiği doğruluk oranı maksimum %84,5 civarında kalmıştır (Zubrinic, 2018). Bunun yanında Zhao ve Xu'nun detaylarının yukarıda verildiği, 2018'de yayınlanan fikir tespiti (opinion detection) çalışmasında, modellerin doğruluk oranları en iyi %70,02 civarında kalmıştır (Zhao, 2018).…”