Abstract:

The authors of this paper, propose a design of an intelligent approach for mapping short answer with the items /questions using Natural Language Processing (NLP). In the present era of technology there has been a significant impact of automated assessment, most of these uses MCQ (Multiple Choice Question). These assessment technique uses web based teaching and learning system. This technique, if implemented, will facilitate the quick assessment of subjective type answer using technology enabled teaching and learning system. In this research article authors have discussed a method for assessing answers based on NLP. Authors have divided the system into two modes, first mode is belongs to instructor and second one belongs to learner and each mode works separately.

In instructor mode, a set of items/questions are provided to the system using a file. In this mode the instructor also provide the tentative correct answers to the system. In the learner mode, a learner opts for a set of questions based on a particular topic in a related domain. The system provides a set of items from the item bank. The learner provides the answer through the interface for the respective items in given instance of time. Once the learner finishes answering the question, the system evaluates using the proposed algorithm.

The proposed answers submitted by the learner are first pass through the process of tokenization. Next the tokens are passed through POS tagging Stop word removal and stemming process will be applied to the answer; these processes alleviate all prefix, suffix statements attached with answer. Subsequent to this the remaining part is mapped with answer stored in repository. This system provides the flexibility for a learner to write answer in different ways. It also provides flexibility to write answers in their own style provided answer must be grammatically correct in terms of the language.

The authors of this paper used different algorithms to improve this system technically, for POS tagging we have used “The penn tree-bank” algorithm and for stemming process “Porter stemming algorithm” have been used.

Keywords:

e-learning, Natural language processing, POS tagging, Stemming, Tokenization.