| Popis: |
Latent Semantic Analysis (LSA), in general, can be considered as an excellent information retrieval technique, but for this specific task of Descriptive Answer Assessment (DAA) some more explorations are required and it is still considered as an open problem. This paper discusses and evaluates the performance of LSA for DAA, through experimentation and deep investigations. Several state-of-the-art claim that LSA correlates with the human assessor’s way of assessment. With this as background, we investigated assessment of students’ descriptive answer using Latent Semantic Analysis (LSA). In the course of research, it was discovered that LSA has limitations like: LSA research usually involves heterogeneous text (text from various domains) which may include irrelevant terms that are highly susceptible to noisy, missing and inconsistent data. The experiments revealed that the general behavior of LSA has an adverse impact on DAA. It has also been observed that factors which influence the performance of LSA are corpus preprocessing, the creation of term-document matrix with and without weight function and choice of dimensionality. |
