{"id":105775,"date":"2023-02-13T16:17:54","date_gmt":"2023-02-13T15:17:54","guid":{"rendered":"https:\/\/diaridigital.urv.cat\/?p=105775"},"modified":"2023-03-13T16:18:19","modified_gmt":"2023-03-13T15:18:19","slug":"a-linguistic-model-enables-feeling-to-be-evaluated-in-social-media-texts","status":"publish","type":"post","link":"https:\/\/diaridigital.urv.cat\/en\/a-linguistic-model-enables-feeling-to-be-evaluated-in-social-media-texts\/","title":{"rendered":"A linguistic model enables feeling to be evaluated in social media texts"},"content":{"rendered":"

Words that we use <\/span>in <\/span>all sorts of<\/span> field<\/span>s<\/span> on a daily basis<\/span> such as <\/span><\/span>fast, slow, <\/em><\/span>nice<\/span>, <\/span>hot<\/span><\/em><\/span> or<\/em> <\/span><\/span>normal <\/span><\/span><\/em>implicit<\/span>ly bear<\/span> a<\/span>n<\/span> information <\/span>load <\/span>that <\/span>is of <\/span>increasing <\/span>importance<\/span> to companies and organi<\/span>s<\/span>ations. <\/span>E<\/span>valuative language, which is used every day, makes communicative interactions interesting <\/span>because<\/span> it<\/span> provide<\/span>s<\/span> essential information. <\/span>Sentiment<\/span> analysis is an area that has received <\/span>considerable<\/span> attention in recent years due to the massive use of social media<\/span>,<\/span> which<\/span> generate <\/span>large amounts<\/span> of evaluative text <\/span>from<\/span> users <\/span>about<\/span> all kinds of products and services. Given the interest <\/span>aroused<\/span> by the analysis of these texts, a research team from the<\/span> URV\u2019s<\/span> <\/span>Department of Romance <\/span>Studies<\/span><\/span><\/a> has developed a technique that includes <\/span>various<\/span> mathematical and linguistic <\/span>methods<\/span> and <\/span>which can<\/span> formally model evaluative statements and capture or extract <\/span>the <\/span>feeling (or <\/span>e<\/span>valuation) <\/span>underlying<\/span> a wide variety of<\/span> linguistic expressions. The result of their research, which has been done in collaboration with the IRAFM Centre of Excellence in the Czech Republic, has been published in the scientific journal <\/span><\/span>Mathematics<\/span><\/span><\/em>.<\/span><\/span>\u00a0<\/span><\/p>\n

To analyse <\/span>feeling<\/span>, computational tools are used to detect and <\/span>assess<\/span> evaluative<\/span> language, in terms of polarity<\/span>,<\/span> that is<\/span>:<\/span> they automatically classify texts <\/span>in terms of<\/span> the positive or negative connotation of the language used. This analysis attempts to <\/span>determine<\/span> a person’s attitude to a topic. Attitude can be a judg<\/span>e<\/span>ment or evaluation, <\/span>an<\/span> affective state (<\/span>the <\/span>emotional state of the author when writing), or the emotional communicative intention (the emotional effect that the author tries to cause in the reader). <\/span>In order to<\/span> function,<\/span> these <\/span>sentiment<\/span> analysis tools require formal models that can describe evaluative language in <\/span>such <\/span>a way<\/span> that <\/span>it can be processed by <\/span>a machine.<\/span><\/span>\u00a0<\/span><\/p>\n

E<\/span>valuative language is<\/span> said to be<\/span> uncertain<\/span> or vague, since it is <\/span>very difficult<\/span> to delimit the meaning of everyday words such as <\/span><\/span>good<\/em>, bad<\/em>, big<\/em>, small<\/em>, love<\/em>, hate<\/em>, etc.<\/span> <\/span>For example, a 5-year-old can be \u201ctall\u201d if <\/span>they are<\/span> 130\u202fcm, and an adult basketball player, on the other hand, is \u201ctall\u201d if <\/span>they are<\/span> 220 cm. This variability can also be found across cultures: for example, the final meaning of the adjective \u201ctall\u201d is surely different in the American and Japanese conceptions. Although the final meaning is different, everyone can understand that \u201ctall\u201d means <\/span>a <\/span>high value on a height scale.<\/span> A model to characterize this \u201c<\/span>fuzziness<\/span>\u201d in meaning is a diffuse model, and this is the basis of the proposal of this research, headed by Adri\u00e0 Torrens and Mar\u00eda Dolores Jim\u00e9nez, of the Research Group in Mathematical Linguistics <\/span>from<\/span> the <\/span>URV\u2019s <\/span>Department of Romance <\/span>Studies<\/span>, together with Vil\u00e9m Nov\u00e1k, <\/span>from<\/span> the University of Ostrava in the Czech Republic.<\/span><\/span><\/p>\n

\"\"<\/a>
The research team consisting of Adri\u00e0 Torrens and Mar\u00eda Dolores Jim\u00e9nez has carried out this research.<\/figcaption><\/figure>\n

Formally <\/span>modelling<\/span> evaluative statements and capturing or extracting the feeling (or <\/span>e<\/span>valuation) behind these linguistic expressions is undoubtedly a <\/span>challenge. Typically, both machine learning algorithms and dictionary techniques (known as \u201cbag of words\u201d) are used for these tasks.<\/span><\/span>\u00a0<\/span><\/p>\n

Learning algorithms focus on aspects of computational performance. In general, these techniques do not <\/span>provide<\/span> sufficient features<\/span> from the point of view of linguistic processes. This research aims to present <\/span>a new approach<\/span> based on a formal interdisciplinary model that <\/span>identifies<\/span> and analyses the <\/span>uncertain<\/span> nature and vague information of evaluative expressions, addressing many of their nuances and offering an \u201cexplanatory\u201d idiosyncrasy.<\/span><\/span>\u00a0<\/span><\/p>\n

The model proposed by this research team combines a property grammar and a fuzzy logic model. The <\/span>property <\/span>grammar <\/span>establishes<\/span> the constraints\/<\/span> conditions that a linguistic structure must meet to be <\/span>appropriate<\/span>. <\/span>The fuzzy model <\/span>captures<\/span> the vagueness of these expressions (\u201ctall\/high\u201d can mean 130 cm or 220 cm<\/span>), <\/span>and<\/span> determine<\/span>s<\/span> the degree of positivity and\/or negativity of an expression (any word can be more or less positive or negative depending on the context in which it is used)<\/span>.<\/span> This model is expected to have <\/span>numerous<\/span> applications and a major impact <\/span>on<\/span> areas such as data mining, language learning tools, automatic authorship detect<\/span>ion<\/span>, etc.<\/span><\/span>\u00a0<\/span><\/p>\n

The <\/span>re<\/span>search <\/span>does not end here.<\/span> According to the research team, the next step is to carry out an interdisciplinary project with professionals in psychology, computational engineering, <\/span>linguistics<\/span> and lexicography <\/span>to<\/span> construct a set of evaluative nuclei that can be applied in sentiment analysis, <\/span>similar to<\/span> the WordNet project at Princeton University. \u201cThis would help to <\/span>identify<\/span> violent <\/span>language and<\/span> would also have benefits for data analysis in <\/span>such<\/span> service sector<\/span>s as<\/span> tourism<\/span>.<\/span> It would also help<\/span> detect cognitive problems in relation to the semantic level of language,\u201d explains Mar\u00eda Dolores Jim\u00e9nez, one of the researchers involved in this research.<\/span><\/span>\u00a0<\/span><\/p>\n

Reference<\/strong>: Torrens-Urrutia<\/span><\/span>, A., <\/span>Nov\u00e1k<\/span>, V., & Jim\u00e9nez-L\u00f3pez, M. D. (2022). <\/span><\/span>Describing Linguistic Vagueness of Evaluative Expressions Using Fuzzy Natural Logic and Linguistic Constraints. <\/span><\/span>Mathematics<\/span><\/span><\/em>, <\/span><\/span>10<\/span><\/span> (15), 2760. <\/span><\/span>https:\/\/www.mdpi.com\/2227-7390\/10\/15\/2760<\/span><\/span><\/a>\u00a0<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"

A research team from the URV has developed a system that identifies feelings in evaluative language. The results of their research have been published in the scientific journal Mathematics<\/p>\n","protected":false},"author":10,"featured_media":104753,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[806,3486,789,818,195,4244,178],"tags":[],"class_list":["post-105775","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-communicating-science","category-romance-studies","category-faculty-arts","category-press-releases","category-research","category-groups","category-science-technology"],"acf":[],"_links":{"self":[{"href":"https:\/\/diaridigital.urv.cat\/en\/wp-json\/wp\/v2\/posts\/105775","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/diaridigital.urv.cat\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/diaridigital.urv.cat\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/diaridigital.urv.cat\/en\/wp-json\/wp\/v2\/users\/10"}],"replies":[{"embeddable":true,"href":"https:\/\/diaridigital.urv.cat\/en\/wp-json\/wp\/v2\/comments?post=105775"}],"version-history":[{"count":0,"href":"https:\/\/diaridigital.urv.cat\/en\/wp-json\/wp\/v2\/posts\/105775\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/diaridigital.urv.cat\/en\/wp-json\/wp\/v2\/media\/104753"}],"wp:attachment":[{"href":"https:\/\/diaridigital.urv.cat\/en\/wp-json\/wp\/v2\/media?parent=105775"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/diaridigital.urv.cat\/en\/wp-json\/wp\/v2\/categories?post=105775"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/diaridigital.urv.cat\/en\/wp-json\/wp\/v2\/tags?post=105775"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}