Towards automatic assessment of argumentation in theses justifications. García-Gorrostieta, J., López-López, A., & González-López, S. Volume 10474 LNCS , 2017.
abstract   bibtex   
© Springer International Publishing AG 2017. Argumentation during the academic life is a critical skill when writing. This skill is needed to communicate clearly ideas and to convince the reader of the presented claims. However, not many students are good arguers and this is a skill difficult to master. This paper presents advances in the development of an argument assessment module. Such module supports students to identify argumentative paragraphs and determine the level of argumentation in the text. The task is achieved employing machine learning techniques with lexical features such as unigrams, bigrams, and argumentative markers categories. We based the module on an annotated collection of student writings, that serves for training. We performed an initial experiment to evaluate argumentative paragraph identification in the justification section of theses, reaching encouraging results, when compared against previously proposed approaches. The module is one component of a Thesis Writing Tutor, an Internet-based learning software for academic writing.
@book{
 title = {Towards automatic assessment of argumentation in theses justifications},
 type = {book},
 year = {2017},
 source = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)},
 identifiers = {[object Object]},
 keywords = {Academic writing,Argumentation studies,Computer-assisted argument analysis,Corpus analysis,Intelligent tutoring system},
 volume = {10474 LNCS},
 id = {06303cc5-57dc-30ff-96e5-a6d3192ae5e5},
 created = {2017-09-29T11:55:51.290Z},
 file_attached = {false},
 profile_id = {ff161cb2-cc48-3995-9f7c-47863e644f60},
 last_modified = {2017-09-29T11:55:51.290Z},
 read = {false},
 starred = {false},
 authored = {true},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {© Springer International Publishing AG 2017. Argumentation during the academic life is a critical skill when writing. This skill is needed to communicate clearly ideas and to convince the reader of the presented claims. However, not many students are good arguers and this is a skill difficult to master. This paper presents advances in the development of an argument assessment module. Such module supports students to identify argumentative paragraphs and determine the level of argumentation in the text. The task is achieved employing machine learning techniques with lexical features such as unigrams, bigrams, and argumentative markers categories. We based the module on an annotated collection of student writings, that serves for training. We performed an initial experiment to evaluate argumentative paragraph identification in the justification section of theses, reaching encouraging results, when compared against previously proposed approaches. The module is one component of a Thesis Writing Tutor, an Internet-based learning software for academic writing.},
 bibtype = {book},
 author = {García-Gorrostieta, J.M. and López-López, A. and González-López, S.}
}

Downloads: 0