=Paper=
{{Paper
|id=Vol-2354/w2preface
|storemode=property
|title=None
|pdfUrl=https://ceur-ws.org/Vol-2354/w2preface.pdf
|volume=Vol-2354
|dblpUrl=https://dblp.org/rec/conf/its/BelandDL18
}}
==None==
<pdf width="1500px">https://ceur-ws.org/Vol-2354/w2preface.pdf</pdf>
<pre>
     Learning analytics
workshop: Building bridges
 between the Education and
the Computing communities

 Sébastien Béland 1, Michel C. Desmarais 2, and Nathalie Loye 1

       1 Université de Montréal, 2 Polytechnique Montréal, Canada




                                 51
52
  Learning analytics workshop: Building bridges
   between the Education and the Computing
                   communities

        Sébastien Béland1 , Michel C. Desmarais2 , and Nathalie Loye1
            1
                Université de Montréal, 2 Polytechnique Montreal, Canada

    The Learning Analytics (LA) and Educational Data Mining (EDM) fields
have generated a wealth of research over the last decade, including two yearly
conferences and two scientific journals. However, these topics are relatively new
in the field of educational science. This workshop brings together researchers
and practitioners to share their perspective on how this research has impacted
the education field.
    Labarthe, Luengo, and Bouchet reports on the very topics that Educational
Data Mining and Learning Analytics have addressed in the last decade. Through
the analysis of papers from tens of conferences and journals, they reveal the main
research trends of each field and show their similarities and di↵erences.
    Two other workshop papers describe practical applications of LA techniques
over typical problems faced by educational practitioners. Xu, Chen, and Wu de-
scribe the results of a Neural Network approach to predict honor student grades
from a wide diversity of factors, ranging from Internet usage to past grades.
They show that the Neural network technique can achieve substantially better
accuracy than more traditional linear regression methods, giving weight to the
advantages of machine learning techniques over standard statistical techniques.
Desmarais addresses the problem of selecting candidates for limited admission
programs, when candidate sources have no common grading schemes. He shows
that, using statistical distribution assumptions combined with an optimization
technique and historical scores from the host institution, the proposed approach
can improve the expected score of accepted students by about one third standard
deviation.
    The workshop also hosted demos of innovative Learning Analytics tools to
help college administrators in their tasks of reporting performance indicators
and provide insights to guide the creation, refinement and evolution of study
programs, and presentations of educational games analytics.




                                          53

</pre>