Automatic Content Assessment of Online Complaint Applications

• Lucas G. S. Félix UFSJ
• João Victor Silveira UFSJ
• Washington Luiz UFMG
• Diego Dias UFSJ
• Leonardo Rocha UFSJ

DOI: https://doi.org/10.5753/kdmile.2018.27384

Abstract

The Internet has experienced a notable expansion and popularization in recent years. It is estimated that by the year 2020 there will be about 40 trillion gigabytes of data generated. There are several scenarios where new techniques and methodologies have been proposed so that relevant information can be extracted from this large volume of data. A recent example are online complaints applications, such as ReclameAqui, which act as a spokesperson for dissatisfied consumers who report their bad experiences with certain products and/or services. This data can represent a rich source of information that can be used by companies to improve it. In this work, we propose a methodology that, through the combination of topic modeling techniques and sentiment analysis, is able to extract from this data useful information, rich in details, that can help companies to identify problems more consistently and quickly. in products and services. We evaluated our methodology with a collection of comments collected from the ReclameAqui application, another one from Twitter and another one from PROCON, all of them related to the four largest telephone companies in Brazil (TIM, OI, VIVO and CLARO). In our evaluations, we have shown that the wealth of details that can be extracted from ReclameAqui and Twitter are much greater when compared to those registered in PROCON. In addition, demonstrating that, as it is an extremely informal application, extracting information from Twitter requires more computational and human resources, which makes online complaints application comments the best alternative to extract useful information.

Keywords: Topic Modeling, Sentiment Analysis, Internet-Based Applications

References

Almeida, T. G., Souza, B. A., Menezes, A. A., Figueiredo, C., and Nakamura, E. F. Sentiment analysis of portuguese comments from foursquare. In Proceedings of the 22nd Brazilian Symposium on Multimedia and the Web. ACM, pp. 355–358, 2016.

Auden, W. H. The Complete Works of W. H. Auden: Prose. Vol. 2. Princeton University Press, 2002.

Cheng, X., Yan, X., Lan, Y., and Guo, J. Btm: Topic modeling over short texts. IEEE Transactions on Knowledge and Data Engineering 26 (12): 2928–2941, 2014.

Gilbert, C. H. E. Vader: A parsimonious rule-based model for sentiment analysis of social media text. In Eighth International Conference on Weblogs and Social Media (ICWSM-14). Available at (20/04/16) [link], 2014.

Hong, L. and Davison, B. D. Empirical study of topic modeling in twitter. In Proceedings of the first workshop on social media analytics. ACM, pp. 80–88, 2010.

Jankowski-Lorek, M. and Zieliński, K. Document controversy classification based on the wikipedia category structure. Computer Science vol. 16, 2015.

Luiz, W., Viegas, F., Alencar, R., Mourão, F., Salles, T., Carvalho, D., Gonçalves, M. A., and Rocha, L. A feature-oriented sentiment rating for mobile app reviews. In Proceedings of the 2018 World Wide Web Conference. WWW ’18. International World Wide Web Conferences Steering Committee, Republic and Canton of Geneva, Switzerland, pp. 1909–1918, 2018.

Pak, A. and Paroubek, P. Twitter as a corpus for sentiment analysis and opinion mining. In LREc. Vol. 10, 2010.

Rocha, L., Mourão, F., Silveira, T., Chaves, R., Sa, G., Teixeira, F., Vieira, R., and Ferreira, R. Saci: Sentiment analysis by collective inspection on social media content. Web Semantics: Science, Services and Agents on the World Wide Web vol. 34, pp. 27–39, 2015.

Sá, G., Silveira, T., Chaves, R., Teixeira, F., Mourão, F., and Rocha, L. Legi: Context-aware lexicon consolidation by graph inspection. In Proceedings of the 29th Annual ACM Symposium on Applied Computing. ACM, pp. 302–307, 2014.

Zhao, W. X., Jiang, J., Weng, J., He, J., Lim, E.-P., Yan, H., and Li, X. Comparing twitter and traditional media using topic models. In European Conference on Information Retrieval. Springer, pp. 338–349, 2011.