responses, but also does not take into account the order of words and also cannot deal
with lexical variability.
The technique of extracting information consists in structuring information from
the text. Can be used to extract communication between concepts. First, the text is
divided into concepts and their relationships. Then the dependencies found are
compared with the expert data (introduced by the instructors) to give the student an
assessment.
Full Natural Language Processing (NLP) involves a parsing analysis of the text
and finds the meaning of the student's response and compares with the teacher's answer
for assigning a final score.
Methods based on ontology, keyword analysis, natural language processing and
information mining techniques are the main approaches to evaluating the text. Full text
parsing and semantic analysis are difficult to implement and very difficult to localize
into other languages. Thus, the extraction of information offers an affordable and more
reliable approach, using the NLP tools to search for specific content in the text and
without in-depth analysis.
Advance NLP method evaluates the text, calculating the score, based on a clear
correlation between the concepts between the student's response and the teacher's
answer (i.e., an expert response) [11]. If there is access to more than one answer, the
similarity is evaluated with respect to each response independently, and the best is used
to determine the final score. The text is divided into tokens (words, numbers, and
punctuation symbols), the boundaries of sentences are indicated. Conducted
synonymous, numeric (for example, '7th' equates to 'seventh'), acronymic
transformation. Concepts turn into intermediate forms and agree on the basis of several
modules that compare essence and heuristic representation of concepts. On the basis of
work of all modules an estimation is made.
CONCLUSION
The article describes all details and functional implementation of system that can
enhance education in any university or school. System itself has some relation to social-
network frameworks that are used commonly by profitable organizations, but have
some security approaches to avoid exploit of the system for profit or other non-
educational purposes. Some management, and motivational functionality mentioned.
As part of the system, article describes details on evaluation method that allows to have
statistically correct scores for each question in the exam bundle. Each type of questions
(“yes-no”, multiple choice, etc.) is described from algorithmically point of view as well
as recalculation algorithm, that keeps question scores in accordance to students
knowledge. Huge attention focused on possible NLP usage for open text questions.
This approach would hugely impact educational system once developed. Natural
- 509 -