Publication: Herramienta de marcado objetiva y subjetiva para la evaluación de oradores
Loading...
Official URL
Full text at PDC
Publication Date
2018
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
Hablar en público es algo cada vez más común en nuestras vidas. Con el objetivo de crear herramientas automáticas que puedan enseñar a hablar en público, es necesario capturar los requisitos del dominio al que nos enfrentamos. A pesar de que es algo cada vez más cotidiano, no hemos encontrado, en toda la literatura consultada, ninguna información útil para la creación de herramientas de este tipo.
Antes de que un ordenador pueda dar feedback a un humano que le ayude a fortalecer sus virtudes y a corregir sus errores como orador, es necesario saber las acciones que son correctas o incorrectas frente a un público. Este proyecto está enfocado a la captura de los requisitos del buen y del mal orador.
La manera óptima que hemos encontrado para capturar estos requisitos ha sido mediante la creación de una herramienta en la que audiencias reales valoren a los oradores. Dichas valoraciones pueden ir desde no mirar al público hasta el uso excesivo de muletillas. De esta manera, hemos establecido las bases para la creación de un lenguaje de marcado que refleje lo que debe hacer el orador para conseguir realizar una buena presentación. Además, con el objetivo de conocer la opinión subjetiva de la audiencia, se ha incluido, en la herramienta desarrollada, la posibilidad de opinar sobre determinados videos. Esta herramienta se utilizó en un experimento para obtener datos más fiables sobre la efectividad de la misma, en relación con el objetivo que perseguía este trabajo.
Finalmente, con los datos que los participantes del experimento nos facilitaron, se buscó la relación entre las acciones del orador y la valoración de la audiencia, buscando demostrar cómo lo que hace el orador afecta en gran medida a la opinión que se tiene sobre su presentación.
Generally, giving a speech is something common in everyone’s life. With the aim of creating automatic tools that can teach to speak in public, it is required to capture the requirements of the domain we are facing. In spite of it is something necessary, we have not found any useful information for creating this tool. First of all, it is necessary to capture the actions that are right or wrong in front of an audience and collect the results. Using them, a computer could give feedback to a human to help him to develop his oral skills, improving his strengths and overcoming his weaknesses as a speaker. This project focuses on capturing the requirements of a good or a bad speaker. The best way that we have found to capture these requirements is by creating a tool in which real audiences could evaluate speakers. These assessments can range from looking at the public to the excessive use of pet words. In this way, we seek to set up the basis to create a markup language that let us know what the speaker should do to get a good presentation. In addition, with the aim of knowing the audience’s personal opinion, we added the possibility of reviewing certain videos is implemented as well. In regards to the aim of this project, this tool was used in a guided experiment to obtain more accurate and reliable data and to check the overall effectiveness of our tool. In conclusion, using the data provided by the participants of the experiment, we have searched the relationship between the speaker's actions and the audience's assessment, seeking to demonstrate how the speaker greatly affects the opinion that he has about his presentation.
Generally, giving a speech is something common in everyone’s life. With the aim of creating automatic tools that can teach to speak in public, it is required to capture the requirements of the domain we are facing. In spite of it is something necessary, we have not found any useful information for creating this tool. First of all, it is necessary to capture the actions that are right or wrong in front of an audience and collect the results. Using them, a computer could give feedback to a human to help him to develop his oral skills, improving his strengths and overcoming his weaknesses as a speaker. This project focuses on capturing the requirements of a good or a bad speaker. The best way that we have found to capture these requirements is by creating a tool in which real audiences could evaluate speakers. These assessments can range from looking at the public to the excessive use of pet words. In this way, we seek to set up the basis to create a markup language that let us know what the speaker should do to get a good presentation. In addition, with the aim of knowing the audience’s personal opinion, we added the possibility of reviewing certain videos is implemented as well. In regards to the aim of this project, this tool was used in a guided experiment to obtain more accurate and reliable data and to check the overall effectiveness of our tool. In conclusion, using the data provided by the participants of the experiment, we have searched the relationship between the speaker's actions and the audience's assessment, seeking to demonstrate how the speaker greatly affects the opinion that he has about his presentation.
Description
Máster en Ingeniería Informática, Facultad de Informática, Departamento de Ingeniería del Software e Inteligencia Artificial, curso 2017-2018.