Developing Aura, as a social media over sentiment analysis context, implies to consider some relevant aspects, such as the advantages and challenges of social media and sentiment analysis.
There are a rising number of social media offering popular applications to hundreds of millions of active users. Social networks range from social communities and discussion groups to recommendation engines, tagging systems, mobile social networks, games, and virtual worlds, these websites have modified the way users communicate and navigate on the Internet [1].
Applying sentiment analysis in Aura application can be useful in several ways because there are some advantages of sentiment analysis for diverse areas of our society [2].Ā Ā For example, this subject helps the decision-making in marketing for the success of an advertising campaign or a new product launch. Also, it can define a popular product or service and its particular characteristics from demographics like or dislike.
The Social Media usage is growing and is driven by these challenges:
- How can a user be heard?
- What kind of information source should a user use?
- How can the user experience be improved?
Answers to these questions may be found in the social media data. These challenges present wide opportunities for companies to develop new algorithms and methods for social media [3].
Understanding peopleās sentiments
The high volume of user-generated content that is created on social media sites every day is an important factor. Thus, Aura development would consider this trend and continue with exponentially more content in the future. The challenge would be critical to address management and utility of massive user-generated data.
The amount of user-generated data in Aura would be a source for a range of decision-making process in different areas. This is because these data can have a useful resource: understanding the sentiments of people. There is a recognition that understanding peopleās feelings may be helpful in identifying the others’ problems and strategies strengths [4].
However, sentiment analysis presents many challenges. The first one is regarding ambiguity ā in one situation, a word could be considered positive and at the same time be considered negative in another situation. A second challenge is related to different ways of expressing opinions – people don’t always express opinions in the same way. For example, in applications like Twitter or blogs, people share different opinions in the same sentence which is easy for a human to understand, but more difficult for a computer to reasoning [2].
What are the advantages of using information about sentiments?
The use of this information can be applied to make wiser decisions related to the use of resources, to make improvements in organizations, providing better products/services, and ultimately to improve the citizen lifestyle and the human relations in order to achieve a better society [4]. An example of this application is the impact of tracking peopleās feelings on products, services and events, which allow enterprise managers to have knowledge and parameters to decision-making. Another example is city council administrators that could have the opportunity for improving the services offered to citizens and for addressing challenges of development and sustainability more efficiently based on what people feel [4].
Social media is the current environment for data collection and analysis of sentiments of people. People can share and comment on everything, from personal thoughts to common events or topics in society. The access to social media also can provide more information in the form of hidden metadata. For instance, Operating System language, device type, capture time and geographical location [4].
What are the disadvantages of automatic sentiment analysis?
Despite the possible positive outcomes shown, there are some disadvantages in applying automatic analysis due to the difficulty to implement it because of the ambiguity of natural language and also the characteristics of the posted content. The analysis of tweets is an example of this, for they are usually coupled with hashtags, emoticons and links, creating difficulties in determining the expressed sentiment. In addition, there is a need for automatic techniques that require large datasets of annotated posts or lexical databases where emotional words are associated with sentiment values. Another important aspect is that analyses are suitable for the English language, in which there is a limitation for other languages [4].
In the field of sentiment analysis are some challenges in a range of scenarios, in terms of architecture and application domains with unclear or scarce datasets. Also, there is a lack of labelled data, which can pose a barrier to the advancements in this area [5].
References
[1] Faloutsos, M., Karagiannis, T. & Moon, S. (2010) Online social networks. IEEE Network. [Online] 24 (5), 4ā5. Available from: doi:10.1109/MNET.2010.5578911.
[2] Vinodhini, G. & Chandrasekaran, RM. (2012) Sentiment Analysis and Opinion Mining: A Survey. International Journal of Advanced Research in Computer Science and Software Engineering. Available from: https://pdfs.semanticscholar.org/8aba/3210964e44b6b1aef5c472a3ae9671d24cf8.pdf
[3] Gundecha, P. & Liu, H. (2012) Mining Social Media: A Brief Introduction. In: 2012 TutORials in Operations Research. [Online]. INFORMS. pp. 1ā17. Available from: doi:10.1287/educ.1120.0105 [Accessed: 10 April 2018].
[4] Furini, M & Montangero, M. (2016) TSentiment: On gamifying Twitter sentiment analysis – IEEE Conference Publication. [Online]. Available from: http://ieeexplore.ieee.org/abstract/document/7543720/ [Accessed: 27 March 2018].
[5] Cirqueira, D., VinĆcius, L., Pinheiro, M., Jacob Junior, A., et al. (2017) Opinion Label: A Gamified Crowdsourcing System for Sentiment Analysis Annotation. In: 16 November 2017 p. [Accessed: 27 March 2018].