Probabilistic model of knowledge assessment

Kenpei Shiina*, Shinichi Asakawa, Shigeki Sakata, Nobuko Takahashi, Tatsuro Makino

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


A new latent trait model, which has the properties of both IRT (Item Response Theory) and Latent Class Model, is proposed. An important new feature of this model is that it introduces partial orders into 'latent classes' of Latent Class Model and this amounts to the replacement of the latent linear structure used in IRT with a latent partially ordered structure. Representation by partially ordered structure has several advantage over traditional representation models and an application to the test items of elementary statistics revealed the effectiveness of the model.

Original languageEnglish
Pages (from-to)27-36
Number of pages10
JournalBulletin of Centre for Informatics, Waseda University
Publication statusPublished - 1991 Jul 1

ASJC Scopus subject areas

  • Engineering(all)


Dive into the research topics of 'Probabilistic model of knowledge assessment'. Together they form a unique fingerprint.

Cite this