Comparison Method for Emotion Detection of Twitter Users

The development of technology has enabled the use of new ways and methods to determine the emotion of sharing on social media. For areas such as media and advertising, social media plays an important role today. In this study, Latent Dirichlet Allocation (LDA) and Non-Negative Matrix Factorization (NMF) methods in subject modeling were used to determine the emotions of tweets thrown through Twitter. In addition, the model was also supported by an LDA-based method to increase the success of the system. The dataset consists of 5 emotions; angry, fear, happy, sadness and surprised. The success of all topic modeling methods used in the study was measured and most successful method was NMF. Then, success of the machine learning algorithms were measured by creating file according to Weka with word weights and class label of the topics. The most successful method was n-stage LDA while the most successful algorithm was Random Forest.

[1]  Banu Diri,et al.  Classification of TurkishTweet emotions by n- stage Latent Dirichlet Allocation , 2018, 2018 Electric Electronics, Computer Science, Biomedical Engineerings' Meeting (EBBT).

[2]  Olfa Nasraoui,et al.  Distributed LDA-based Topic Modeling and Topic Agglomeration in a Latent Space , 2014, SNOW-DC@WWW.

[3]  Hui Zhang,et al.  Experimental explorations on short text topic mining between LDA and NMF based Schemes , 2019, Knowl. Based Syst..

[4]  H. Sebastian Seung,et al.  Algorithms for Non-negative Matrix Factorization , 2000, NIPS.

[5]  Loni Hagen,et al.  Content analysis of e-petitions with topic modeling: How to train and evaluate LDA models? , 2018, Inf. Process. Manag..

[6]  N. D. du Preez,et al.  Leveraging Unstructured Information Using Topic Modelling , 2008 .

[7]  Derek Greene,et al.  An analysis of the coherence of descriptors in topic modeling , 2015, Expert Syst. Appl..

[8]  Yuxin Chen,et al.  An experimental comparison between NMF and LDA for active cross-situational object-word learning , 2016, 2016 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob).

[9]  V. P. Pauca,et al.  Nonnegative matrix factorization for spectral data analysis , 2006 .

[10]  Michael I. Jordan,et al.  Latent Dirichlet Allocation , 2001, J. Mach. Learn. Res..

[11]  Mark Johnson,et al.  More Efficient Topic Modelling Through a Noun Only Approach , 2015, ALTA.

[12]  Banu Diri,et al.  Classification of New Titles by Two Stage Latent Dirichlet Allocation , 2018, 2018 Innovations in Intelligent Systems and Applications Conference (ASYU).

[13]  Nihar Ranjan Roy,et al.  Comparison between LDA & NMF for event-detection from large text stream data , 2017, 2017 3rd International Conference on Computational Intelligence & Communication Technology (CICT).

[14]  Derek Greene,et al.  Stability of topic modeling via matrix factorization , 2017, Expert Syst. Appl..

[15]  Jaegul Choo,et al.  Nonnegative Matrix Factorization for Interactive Topic Modeling and Document Clustering , 2014 .

[16]  David Buttler,et al.  Exploring Topic Coherence over Many Models and Many Topics , 2012, EMNLP.

[17]  Shun-ichi Amari,et al.  A new discriminant NMF algorithm and its application to the extraction of subtle emotional differences in speech , 2012, Cognitive Neurodynamics.