TITLE

Effect of tuned parameters on an LSA multiple choice questions answering model

AUTHOR(S)
LIFCHITZ, ALAIN; JHEAN-LAROSE, SANDRA; DENHiÈRE, GUY
PUB. DATE
November 2009
SOURCE
Behavior Research Methods;Nov2009, Vol. 41 Issue 4, p1201
SOURCE TYPE
Academic Journal
DOC. TYPE
Article
ABSTRACT
This article presents the current state of a work in progress, whose objective is to better understand the effects of factors that significantly influence the performance of latent semantic analysis (LSA). A difficult task, which consisted of answering (French) biology multiple choice questions, was used to test the semantic properties of the truncated singular space and to study the relative influence of the main parameters. A dedicated software was designed to fine-tune the LSA semantic space for the multiple choice questions task. With optimal parameters, the performances of our simple model were quite surprisingly equal or superior to those of seventh- and eighth-grade students. This indicates that semantic spaces were quite good despite their low dimensions and the small sizes of the training data sets. In addition, we present an original entropy global weighting of the answers' terms for each of the multiple choice questions, which was necessary to achieve the model's success.
ACCESSION #
45661812

 

Related Articles

  • Using latent semantic analysis to grade brief summaries: A study exploring texts at different academic levels. Olmos, Ricardo; León, José A.; Jorge-Botana, Guillermo; Escudero, Inmaculada // Literary & Linguistic Computing;Sep2013, Vol. 28 Issue 3, p388 

    In this study, we propose an integrated method to automatically evaluate very brief summaries (around 50 words) using the computational tool latent semantic analysis (LSA). The method proposed is based on a regression equation calculated with a corpus of a 100 summaries (the training sample) and...

  • Computer Aided Analysis of Multiple Choice Test Results. ERGÜN, Ertugrul; AYDIN, Ali // Participatory Educational Research;2015 Special Issue II, p110 

    One of the most widely used assessment technique in educational institutions are the multiple-choice tests. Several analyses have to be made in order to determine the validity and reliability of these multiple-choice tests and items in the test. In order to make some comments about multiple...

  • Feature Selection and Clustering Approaches to the KNN Text Categorization. Gayathri, K.; Marimuthu, A. // International Journal of Computer Science Engineering & Technolo;Sep2012, Vol. 2 Issue 9, p1406 

    Automatic text classification is a discipline at the cross roads of information retrieval machine learning and computational linguistics and consists in the realization of text classifiers. (ie) software systems capable of assigning text to one or more categories or classes, from a pre-defined...

  • The Nature of Automated Essay Scoring Feedback. Dikli, Semire // CALICO Journal;Sep2010, Vol. 28 Issue 1, p99 

    The article focuses on the automated essay scoring (AES), a computer system used for evaluating and scoring essays. It states that AES depends on several machine-learning methods including artificial intelligence (AI), natural language processing (NLP) and latent semantic analysis (LSA). It...

  • Experiences and Advancements from One Year of Explorative Application of an Integrated Model- Based Development Technique Using C&C²- A in SysML. Zingel, C.; Albers, A.; Matthiesen, S.; Maletz, M. // IAENG International Journal of Computer Science;2012, Vol. 39 Issue 2, p165 

    The challenge of uncertainty and ambiguity is ubiquitous in the development of complex systems and needs to be faced. The all-embracing integration of specialists from multiple disciplines is proven to be a major challenge in the product engineering process. This article presents the experiences...

  • Visualizing the software system towards identifying the topic from source code using semantic clustering. Sharma, Kanchan; Brar, Amanpreet Singh // International Journal of Advanced Computer Research;Mar2014, Vol. 4 Issue 14, p350 

    In software re-engineering, domain knowledge are valuable source of information for developers. Here, we describe how the coding standards are helpful for the identification of domain while writing the source code. Internal comments and logical identifier names in source code are the key source...

  • LSAfun - An R package for computations based on Latent Semantic Analysis. Günther, Fritz; Dudschig, Carolin; Kaup, Barbara // Behavior Research Methods;Dec2015, Vol. 47 Issue 4, p930 

    In this article, the R package LSAfun is presented. This package enables a variety of functions and computations based on Vector Semantic Models such as Latent Semantic Analysis (LSA) Landauer, Foltz and Laham ( Discourse Processes 25:259-284, ), which are procedures to obtain a high-dimensional...

  • Automated Essay Scoring Using Incremental Latent Semantic Analysis. Mingqing Zhang; Shudong Hao; Yanyan Xu; Dengfeng Ke; Hengli Peng // Journal of Software (1796217X);Feb2014, Vol. 9 Issue 2, p429 

    Writing has been increasingly regarded by the testers of language tests as an important indicator to assess the language skill of testees. As such tests become more and more popular and the number of testees becomes larger, it is a huge task to score so many essays by raters. So far, many...

  • UML models change impact analysis using a text similarity technique. Kchaou, Dhikra; Bouassida, Nadia; Ben-Abdallah, Hanêne // IET Software;2017, Vol. 11 Issue 1, p27 

    Given the inevitable software evolution, change impact analysis (CIA) is a vital activity in the software development life cycle. Existing CIA methods either focus on one model produced during one development phase or ignore the semantic dependencies among the various models produced throughout...

Share

Read the Article

Courtesy of THE LIBRARY OF VIRGINIA

Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics