CAPITULO II CONDICIONES DE IMPLANTACIÓN DE LOS USOS Y/O
SECCION 4 C. EXPLOTACIÓN DE RECURSO PRIMARIOS
Opinions are central to almost all human activity, as they may influence the behaviour of others. If we need to make a decision, we are interested in other people’s opin- ions. As individuals, for example, we want to know the opinions of existing users of a product before purchasing it, or others’ opinions about political candidates before mak- ing a voting decision in an election. In business, companies and organisations acquire public opinions for marketing purposes, public relations, and political campaigning [113].
In the past, when an individual needed an opinion, they would ask friends and family members. When a company or organisation needed public or consumer opinions, it conducted surveys, opinion polls, and focus groups. With the proliferation of the Web 2.0, there came textual reviews, forum discussions, blogs, micro-blogs, comments, and postings on social networking sites, all generated by us, the users. These forms of me- dia, collectively referred to as user-generated content, have produced an important shift
in the way in which people communicate and share textual knowledge and information, which influence social, political, and economic behaviour worldwide [133].
Sentiments, evaluations, and reviews are becoming very much evident with the grow- ing interest in e-commerce, which is a prominent source of expressing and analysing opinions. Nowadays, customers on e-commerce sites, such as Amazon1, mostly rely on reviews posted by other customers. For instance, in the Review Example (see page 7), we can be satisfied in deciding on purchasing the headphones, following a review describing a customer’s positive experience in using the product.
In business, companies have increasingly realised that consumer voices influence and shape the opinions of other consumers and, ultimately, their brand loyalties and pur- chase decisions. In turn, customers’ opinions are analysed for market research. For example, if a product seller is receiving negative feedback, in the interest of their repu- tation, the service provider might remove or suspend the seller. As well as product reviews, the web abounds with reviews for all kinds of services, ranging from restaur- ants (e.g. Yelp2), holiday destinations (e.g. Trip Advisor3), and films (e.g. IMDb4and
Rotten Tomatoes5).
Online media and social networking sites (e.g. Facebook6 and Twitter7) are also used for public self-disclosure. These open platforms provide and encourage users to in- teractively express their thoughts on global and personal issues across their social net- work, in real time. Most recently, users are expressing their opinions and sharing their experiences regarding the services they use and the products they buy on such social media sites. Companies ::::::Large:::::::::corporate::::::::::::companies:::::such:::as:::::::::Amazon::::and::::the
::::::::::::
supermarket::::::chain::::::Tesco8 have increasingly turned to social media to analyse and re-
1https://www.amazon.co.uk/ 2https://www.yelp.com 3https://www.tripadvisor.co.uk/ 4http://www.imdb.com/ 5https://www.rottentomatoes.com/ 6https://www.facebook.com/ 7https://twitter.com/ 8 :::::::::::::::::::https://www.tesco.com/
spond to consumer opinions.
Aside from short, mostly informal, social posts and product reviews, blogs have also emerged as a platform for expressing self-disclosure. Blogs are frequently updated series of archived posts, typically in reverse-chronological order. They may vary widely in nature and content, and have featured extensively in the popular media, en- tering political campaigns, news organisations, and businesses. As they have grown in popularity, they are often portrayed as online diaries or personal journals, often ex- pressing sentiment towards more pensive subjects, such as health related topics. For example, Claire Greaves [73] openly blogs about her experiences with mental health, a subject which is surrounded by stigma.
Along with news media sites (e.g. BBC News9), such platforms offer a commenting feature, which prompts discussions within online social communities, where users are able to respond and contribute their experiences. This produces a stream of real time opinionated data, which can be used for applications which aim to, for example, predict election outcomes (e.g. [214]) and stock market direction (e.g. [3, 99]).
In its own right, user-generated content represents a rich source of information. The increasing popularity of publishing personal texts suggests that opinionated informa- tion is becoming an important aspect of the textual data on the web [213]. These forms of media serve as a platform to extract heterogeneous opinions that are publicly pub- lished by users from diverse societies. Due to the ever-growing volume of this type of information, the analysis of sentiments and opinions on a large scale is impractical without automatic classification and aggregation.
As a response, sentiment analysis methods and their applications have flourished in recent years. Such applications have spread across several domains, such as politics, finance, and health care [113].
The recent role of social media has produced several studies on the mining of online
political speech [127]. Sentiment analysis has been used to understand voters’ opinions (e.g. [106, 136]), to clarify what public figures support or oppose (e.g. [211]), tracking discussion on political debates (e.g. [43]), predicting election outcomes (e.g. [214]), and determining whether news sources are biased in covering one political party more than another (e.g. [111]). Such applications help enhance the quality of the inform- ation that voters have access to during a political campaign, supporting their voting decisions.
Sentiment analysis has also been utilised in finance. Existing studies in this domain have focused on the relationship between the general mood of the public expressed on social media and the stock market direction (e.g. [21]), the sentiment expressed by expert investors on financial blog sources (e.g. [144]), and the opinions expressed on traditional news and finance media (e.g. [3, 99]). Such applications support financial and economic experts in identifying shifts in the markets. This provides a wider un- derstanding about corporates, and a base to support economic knowledge in decision making, such as risk estimation before making an investment [75].
In health care, understanding a patient’s experience is central to improving the qual- ity of care they may receive in the future. Traditional measures for collating this in- formation include surveys or structured questionnaires which ask specific and limited questions, and are expensive to administer. Patients have begun to report their health care experiences on the web in the form of blogs, social media postings, and health care rating websites. Existing studies in this domain vary, having focused on automat- ically analysing patient opinions and satisfaction ratings for services such as the UK’s National Health Service (e.g. [4, 27]), estimating the sentiment of forum posts written by cancer survivors, as well as studying the changes in these forum members and their sentiments during their treatment [167], and the sentiment expressed in suicide notes (e.g. [196, 158]), which can be used to support online suicide prevention services.