Multitask Learning for Complaint Identification and Sentiment Analysis

In today’s competitive business world, customer service is often at the heart of businesses that can help strengthen their brands. Resolution of customers’ complaints in a timely and efficient manner is key to improving customer satisfaction. Moreover, customers’ complaints play an important role in identifying their requirements which offer a starting point for effective and efficient planning of companies’ overall R&D and new product or service development activities. Having said that, organizations encounter challenges towards automatically identifying complaints buried deep in massive online content. Our current work centers around learning two closely related tasks, viz. complaint identification and sentiment classification. We leverage weak supervision to annotate the corpus with sentiment labels. We propose a deep multitask framework that features a knowledge element that uses AffectiveSpace to infuse commonsense knowledge specific features into the learning process. The framework models complaint identification (the primary task) and sentiment classification (supplementary task) simultaneously. Experimental results show that our proposed multitask system obtains the highest cross-validation accuracy of 83.73 +/- 1.52 % for the complaint identification task and 69.01 +/- 1.74 % for the sentiment classification task. Our proposed multitask system outperforms the single-task systems indicating a strong correlation between sentiment analysis and complaint classification tasks, thus benefiting from each other when learned concurrently.

[1]  Xuanjing Huang,et al.  Adversarial Multi-task Learning for Text Classification , 2017, ACL.

[2]  Heikki Mannila,et al.  Random projection in dimensionality reduction: applications to image and text data , 2001, KDD '01.

[3]  Elite Olshtain,et al.  10. Complaints: A study of speech act behavior among native and non-native speakers of Hebrew , 1987 .

[4]  Virach Sornlertlamvanich,et al.  Recurrent Neural Network with Word Embedding for Complaint Classification , 2016, WLSI/OIAF4HLT@COLING.

[5]  Yoshua Bengio,et al.  Deep Sparse Rectifier Neural Networks , 2011, AISTATS.

[6]  Jason Weston,et al.  End-To-End Memory Networks , 2015, NIPS.

[7]  Gaël Varoquaux,et al.  Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[8]  Erik Cambria,et al.  AffectiveSpace 2: Enabling Affective Intuition for Concept-Level Sentiment Analysis , 2015, AAAI.

[9]  Klaus-Robert Müller,et al.  Efficient BackProp , 2012, Neural Networks: Tricks of the Trade.

[10]  Rich Caruana,et al.  Promoting Poor Features to Supervisors: Some Inputs Work Better as Outputs , 1996, NIPS.

[11]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[12]  Erik Cambria,et al.  Multitask Representation Learning for Multimodal Estimation of Depression Level , 2019, IEEE Intelligent Systems.

[13]  Erik Cambria,et al.  Sentic Computing: Exploitation of Common Sense for the Development of Emotion-Sensitive Systems , 2009, COST 2102 Training School.

[14]  Daniel Jurafsky,et al.  Deconfounded Lexicon Induction for Interpretable Social Science , 2018, NAACL.

[15]  Diana Boxer,et al.  Social distance and speech behavior: The case of indirect complaints , 1993 .

[16]  Sebastian Ruder,et al.  An Overview of Multi-Task Learning in Deep Neural Networks , 2017, ArXiv.

[17]  Abhishek Kumar,et al.  A Multilayer Perceptron based Ensemble Technique for Fine-grained Financial Sentiment Analysis , 2017, EMNLP.

[18]  Erik Cambria,et al.  Affective Computing and Sentiment Analysis , 2016, IEEE Intelligent Systems.

[19]  A. Cohen,et al.  The Production of Speech Acts by EFL Learners , 1993 .

[20]  Erik Cambria,et al.  SenticNet 6: Ensemble Application of Symbolic and Subsymbolic AI for Sentiment Analysis , 2020, CIKM.

[21]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[22]  Marilu B. Rañosa-Madrunio.  The discourse organization of letters of complaint to editors in Philippine English and Singapore English , 2004 .

[23]  A. Trosborg.  Interlanguage Pragmatics: Requests, Complaints, and Apologies , 1994 .

[24]  Marty Laforest.  Scenes of family life - complaining in everyday conversation , 2002 .

[25]  Jeffrey Pennington,et al.  GloVe: Global Vectors for Word Representation , 2014, EMNLP.

[26]  Yurong Liu,et al.  A survey of deep neural network architectures and their applications , 2017, Neurocomputing.

[27]  Bin Wu,et al.  Complaint Classification Using Hybrid-Attention GRU Neural Network , 2019, PAKDD.

[28]  Jun Zhao,et al.  Recurrent Convolutional Neural Networks for Text Classification , 2015, AAAI.

[29]  A. Mahboob,et al.  Models of discourse in the letter of complaint , 2004 .

[30]  Marja Ebba Meinl,et al.  Electronic complaints: an empirical study on British English and German complaints on eBay , 2013 .

[31]  Asif Ekbal,et al.  How Intense Are You? Predicting Intensities of Emotions and Sentiments using Stacked Ensemble [Application Notes] , 2020, IEEE Comput. Intell. Mag..

[32]  Houbing Song,et al.  SentiRelated: A cross-domain sentiment classification algorithm for short texts through sentiment related index , 2018, J. Netw. Comput. Appl..

[33]  Erik Cambria,et al.  Sentic Album: Content-, Concept-, and Context-Based Online Personal Photo Management System , 2012, Cognitive Computation.

[34]  Penelope Brown,et al.  Politeness: Some Universals in Language Usage , 1989 .

[35]  Camilla Vásquez.  Complaints online: The case of TripAdvisor , 2011 .

[36]  Francesco Piazza,et al.  Sentic Web: A New Paradigm for Managing Social Media Affective Information , 2011, Cognitive Computation.

[37]  Diyi Yang,et al.  Hierarchical Attention Networks for Document Classification , 2016, NAACL.

[38]  Yoshua Bengio,et al.  On the Properties of Neural Machine Translation: Encoder–Decoder Approaches , 2014, SSST@EMNLP.

[39]  Ramit Sawhney,et al.  Did you offend me? Classification of Offensive Tweets in Hinglish Language , 2018, ALW.

[40]  Erik Cambria,et al.  Targeted Aspect-Based Sentiment Analysis via Embedding Commonsense Knowledge into an Attentive LSTM , 2018, AAAI.

[41]  Rich Caruana,et al.  Multitask Learning , 1997, Machine Learning.

[42]  Pushpak Bhattacharyya,et al.  Related Tasks can Share! A Multi-task Framework for Affective language , 2020, ArXiv.

[43]  Zhiyuan Liu,et al.  A C-LSTM Neural Network for Text Classification , 2015, ArXiv.

[44]  Xuanjing Huang,et al.  Cached Long Short-Term Memory Neural Networks for Document-Level Sentiment Classification , 2016, EMNLP.

[45]  Yoshua Bengio,et al.  Neural Machine Translation by Jointly Learning to Align and Translate , 2014, ICLR.

[46]  Quan Pan,et al.  Learning Word Representations for Sentiment Analysis , 2017, Cognitive Computation.

[47]  Nikolaos Aletras,et al.  Automatically Identifying Complaints in Social Media , 2019, ACL.

[48]  Tiejun Zhao,et al.  Target-dependent Twitter Sentiment Classification , 2011, ACL.

[49]  Min Yang,et al.  Attention Based LSTM for Target Dependent Sentiment Classification , 2017, AAAI.

[50]  Erik Cambria,et al.  Sentiment and Sarcasm Classification With Multitask Learning , 2019, IEEE Intelligent Systems.

[51]  Xiang Zhang,et al.  Character-level Convolutional Networks for Text Classification , 2015, NIPS.

[52]  Lei Zhang,et al.  A Survey of Opinion Mining and Sentiment Analysis , 2012, Mining Text Data.

[53]  Meng Zhang,et al.  Neural Network Methods for Natural Language Processing , 2017, Computational Linguistics.

[54]  Jian Pei,et al.  Detecting Customer Complaint Escalation with Recurrent Neural Networks and Manually-Engineered Features , 2019, NAACL.

[55]  Eric Gilbert,et al.  VADER: A Parsimonious Rule-Based Model for Sentiment Analysis of Social Media Text , 2014, ICWSM.

[56]  Erik Cambria,et al.  Word Polarity Disambiguation Using Bayesian Model and Opinion-Level Features , 2014, Cognitive Computation.

[57]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.