Pilar Mazo
Logroño, (EFE).- The use of natural language processing systems -artificial intelligence techniques- can help detect, with greater precision, cases of pathological distress, the director of the Data Science Research Group has informed EFE from the International University of La Rioja (UNIR), Luis de la Fuente.
Scientific studies show the relationship between the symptoms associated with anxiety and the characteristics of the language used, he explains.
De la Fuente, together with the doctor in Social Psychology and Behavioral Sciences from UNIR, Joaquín González Cabrera, have designed a project, pending receipt of funding, through which, by combining artificial intelligence with psychological tools, an early detection of possible cases of bullying or cyberbullying that could occur, for example, in a school, always with an approach used by the center’s educational guidance department.
Through the writing of a text on a specific subject, he specified, it is possible to predict with a success rate whether or not the person who has written it suffers from a mental disorder, such as pathological anxiety, he has indicated.
Currently, he has reported, there is no commercial or professional tool to transform the analysis of these texts into data, but there are published scientific studies and knowledge generated to create it.
The pronoun to detect disorders
These studies indicate that different linguistic features, such as the predominance of the use of certain personal pronouns in an essay, may be related to a certain disorder, De la Fuente specified.
“For example, it has been detected that people who suffer from schizophrenia use the third person plural pronoun more frequently, just as those who suffer from obsessive-compulsive disorders predominantly use the first person singular tense and, furthermore, the pronoun ‘I’, which is also the first person singular”.
The use of the pronoun is a “very useful” aspect in this regard, for statistical purposes, because, from natural language processing, that text can be transformed into data, he has detailed.
These data, he explained, indicate the frequency of the use of pronouns in a written text, but, in addition, within this section of use of pronouns, it is also possible to quantify and analyze the frequency of each of the pronominal forms, of the use of verbs, the average length of the sentences used, the number of relative verbs in the sentences used or the association to the emotional state of the words that have been used, among other parameters.
Artificial intelligence opens a path to hope
“All this -he has reported- can be quantified, transformed into data and, once transformed into data, the use of pronouns in people who are diagnosed with a certain disorder is compared with that of the general population”.
It has been observed in scientific studies, he stressed, that “certain disorders are related to the massive use of the first person singular” of the pronoun.
“Unfortunately”, he has indicated, there is a group of disorders that are correlated, but “it is not easy to distinguish which of them is the one that is causing this main use of the pronoun”.
He believes that “artificial intelligence opens up hope for the future to detect cases of pathological distress with greater precision, always with the caution that what is currently being detected is a group of disorders and not a specific one.” EFE
