ISSN (Online) 2581-9429

IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.53 Volume 4, Issue 1, April 2024

Automatic Question Paper Generation and

Weightage Assigning using Bloom’s Taxonomy
Prof. B.S Chaudhary, Janhvi Jagtap, Minakshi Bhavar, Vaishmavi Wagh, Kiran Pal
Department of Computer Engineering
Loknete Gopinathji Munde Institute of Engineering and Research Center, Nashik, India
[email protected], [email protected], [email protected],
[email protected], [email protected]

Abstract: Bloom’s Taxonomy is a classification of learning objectives within education that educators set
for students. The cognitive domain within this taxonomy is designed to verify a student’s cognitive level
during a written examination. Educators may sometimes face the challenge in analysing whether their
examination questions comply within the requirements of the Bloom’s taxonomy at different cognitive
levels. This paper proposes an automated analysis of the exam questions to determine the appropriate
category based on this taxonomy. This rule-based approach applies Natural Language Processing (NLP)
techniques to identify important keywords and verbs, which may assist in the identification of the category
of a question. This work focuses on the computer programming subject domain. At present, a set of 100
questions (70 training set and 30 test set) is used in the research. Preliminary results indicate that the rules
may successfully assist in the identification of the Bloom’s taxonomy category correctly in the exam
questions.

Keywords: NLP, Bloom’s Taxonomy.

I. INTRODUCTION
With the advancement of technology in the education sector, there is a growing need for automating various processes
to make them more efficient and effective. One such process that can benefit from automation is the generation of
question papers for assessments and exams. Bloom's taxonomy is a framework that categorizes different levels of
learning objectives, from basic knowledge recall to higher-order thinking skills like analysis and synthesis. By using
Bloom's taxonomy as a guide, educators can create well-rounded question papers that assess students' understanding
and application of concepts at various levels. Automatic question paper generation using Bloom's taxonomy involves
using software tools to analyze learning objectives, curriculum content, and student performance data to generate
customized question papers tailored to individual student needs.

Blooms taxonomy:

Fig1:- Blooms Taxonomy

Copyright to IJARSCT DOI: 10.48175/568 13

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.53 Volume 4, Issue 1, April 2024

This approach ensures that the assessments are aligned with the learning objectives and provide a comprehensive
evaluation of students' knowledge and skills. By automating the question paper generation process, educators can save
time and effort in creating assessments, while also ensuring that the questions are of high quality and effectively
measure students' learning outcomes. This approach can also help in reducing bias and ensuring fairness in assessments
by providing a standardized and objective method of evaluation. automatic question paper generation using Bloom's
taxonomy is a valuable tool for educators to create meaningful assessments that promote student learning and
achievement. By leveraging technology in this way, educators can enhance the assessment process and provide valuable
insights into students' progress and performance.
Uses advanced algorithms to generate exam questions based on Bloom's Taxonomy that are relevant to the course
content and aligned with learning outcomes. Incorporates a diverse range of question types, such as multiple-choice,
short answer, and essay questions, to ensure the comprehensive evaluation of students' knowledge and skills. Allows
educators to customize and modify generated question papers according to their preferences and curricula. Provides an
intuitive user interface that allows educators to generate question papers efficiently and quickly. Includes a knowledge
base that is updated regularly with relevant and accurate information to generate high-quality questions. Is reliable,
secure, and scalable to handle question papers for various courses, classes, and academic levels. Offers data analysis
and reporting capabilities to enable educators to evaluate the effectiveness of the question papers and the assessment of
student learning. project's scope is to provide an efficient, effective, and flexible solution for generating exam questions
using Bloom's Taxonomy that saves time and enhances the quality of student learning and assessment.

Normalization of the Data:

The process of automatic question paper generation can be normalized using Bloom's taxonomy by ensuring that
questions cover a range of cognitive levels as defined in the taxonomy.
Bloom's taxonomy categorizes cognitive skills into six levels:
1. Remembering: This level involves recalling facts or information.
2. Understanding: This level involves comprehending the meaning of information.
3. Applying: This level involves using knowledge to solve problems or apply concepts.
4. Analyzing: This level involves breaking down information into parts and examining relationships between them.
5. Evaluating: This level involves making judgments or decisions based on criteria and standards.
6. Creating: This highest level involves generating new ideas, products, or ways of viewing things.
To normalize the automatic question paper generation process based on Bloom's taxonomy, each question generated
should be mapped to one of these cognitive levels to ensure a balanced distribution across all levels within the question
paper.
For example, when creating an automatic system for generating questions for a test in mathematics:
- Remembering - A simple multiple-choice question asking students to recall a mathematical formula
- Understanding - A problem-solving scenario that requires students to demonstrate understanding of a concept
- Applying - A real-life application problem that requires students to use mathematical principles in a new context
- Analyzing - A question that asks students to break down a complex problem into its component parts and identify
patterns
- Evaluating - An open-ended question where students must analyze different solutions and justify their choice
- Creating - A challenge where students have to devise an original mathematical problem for their peers
By normalizing the distribution of questions across these cognitive levels, educators can ensure that the assessment
covers various aspects of learning and provides valuable insights into student capabilities beyond simple recall.

Copyright to IJARSCT DOI: 10.48175/568 14

www.ijarsct.co.in
 ISSN (Online) 2581-9429
2581
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access,
Access, Double
Double-Blind, Peer-Reviewed,
Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.53 Volume 4, Issue 1, April 2024

II. METHODOLOGY

Consider the block diagram from Fig1.1

Bloom’s Taxonomy is a classification of learning objectives within education that educators set for students. The
cognitive domain within this taxonomy is designed to verify a student’s cognitive level during a written examexam- ination.
Educators may sometimes face the challenge in analysing whether their examination questions comply within the
requirements of the Bloom’s taxonomy at different cognitive levels. This paper proposes an automated analysis of the
exam questions to determine
rmine the appropriate category based on this taxonomy. This rule rule-based
based approach applies
Natural Language Pro- cessing (NLP) techniques to identify important keywords and verbs, which may assist in the
identification of the category of a question. This work
wor fo- cuses on the computer programming subject domain. At
present, a set of 100 questions (70 training set and 30 test set) is used in the research. Preliminary results indicate that
the rules may successfully assist in the identification of the Bloom’s ttaxonomy
axonomy category correctly in the exam
questions.
Data mining: Data Mining is the computing process of discovering patterns in large data sets involving methods at the
intersection of machine learning, statistics and database systems.
Blooms Taxonomy: Blooms Taxonomy is a classification
clas of the different objectives
jectives and skills that educators set for
their students (learning objectives). The taxonomy was proposed in 1956 by Benjamin Bloom, an educational
psychologist at the University of Chicago.
Natural Language Processing: The field of study that focuses on the interactions tions between human language and
computers is called Natural Language Processing, or NLP for short.
Randomized Algorithm: A randomized algorithm is an algorithm that employs em a degree of randomness as part of its
logic. Formally, the algorithm’s performance will be a random variable determined by the random bits; thus either the
running time, or the output (or both) are random variables.

2.1 Objective
1. To make question papers with varieded questions and which meet learning objectives of the course.
2. To generate the question paper from teacher entered specification within few seconds.
3. To cover all aspects of the course objectives and avoid duplication of questions in the subsequent eexams.

III. LITERATURE SURVEY

The existing Learning Management Systems (LMS) support very basic level or limited tags such as question types .
Even the most preferred LMS, Moodle allows creating only subjective/objective type of questions. Thus automatical
automatically
generating question paper from a teacher’s entered specification using a semantically tagged QR is the need of the hour
today. The system to semi automatically tag the questions of a repository is in place. Here we are proposing a system
which automatically
ally generates the question paper from this semantically tagged question repository. Since the existing
systems are rigid and lack the flexibility of supporting all types of tags, the generated question paper may not be totally
aligned with its given objectives.
tives. Our system supports all four tags and also flexible enough to provide an interface that
allows user to enter specifications for each tag/property in the form of lower and upper bounds. Each property is
specified with a range indicating that value should
should not be lower than minimum value and not exceed the maximum
value of the range. Also, it is rule base system which takes all the combinations of the tags and generates output based

Copyright to IJARSCT DOI: 10.48175/568 15

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.53 Volume 4, Issue 1, April 2024

on the rule applicable. The output is generated in xml format and in word document. A literature survey was started to
understand the need for automatic generation of QP. As mentioned in, many existing LMS support tagging feature but
users may not utilize this feature fully. The comparative study shows that Moodle is best LMs to support large number
of users and also for any educational institution. But it allows user to define only question type. Hence the questions in
the repository may have only basic tags or no tags at all. So it becomes overhead for teachers to tag these questions
before using them. Properly tagged questions can be efficiently retrieved from repository. Hence it is very much
necessary to tag the questions before adding them to repository. Most teachers and researchers recommend four types of
tags such as topic, question type, cognitive level and difficulty level. A system which offers generation of question
paper using user given input parameters considers only fixed range of values. Our system not only supports upper and
lowers bounds for inputs but also supports more granular level of topics than chapters and more question types as
compared to only three types offered by this system. We are using automatically tagged question repository as input
instead of untagged questions. Automatic Question Paper Generator System described in has great interface, but
supports only question type tag. Hence the question paper generated would have only one difficulty level

3.1 Existing Paper-based Systems

The existing system for Question Paper Generation requires human staff to chalk out questions that appear in the
question paper. These teachers or professors select the questions according to the syllabus and pattern as prescribed by
the curriculum. The question paper then may be referred to a higher authority who has the final say in these matters. B.
Limitation of Paper-based Systems As most human working processes, this system suffers due to bias. There might be
some questions which are repeated in many question papers as the professor has a personal inclination towards them.
So there is no guarantee of pure randomly generated question paper. Other problems that may plague this system are
non-availability of staff and resources, natural calamities and accidents. Also, the security of the system can be easily
compromised if leverage over the person responsible for generating question papers is obtained. Other limitations
include: -
a) Lack of storage space
b) Prone to damage
c) Inefficient document transportation
d) Supply costs
e) Poor environmental credentials
f) Limited collaboration
g) Editing problems

3.2 Existing Question Paper Generation Systems

The research paper „Framework for Automatic Examination Paper Generation System‟ has provided a thorough insight
into the process of automated paper generation. As the manual generation of a balanced question paper by an individual
is quite complex, the blending of technology into teaching and learning process is inevitable. A simple and efficient
way for an examination paper generation is provided. A three tier model is provided in this framework. Generation of
Examination Papers is governed by the Syllabus Engine, Pattern Composer and Question Aggregator. The generated
question paper is based on the pattern or skeleton of the course. Another component called Bank Management takes
care of User Rights and Privilege assignment. Questions are entered through the Question Aggregator. The attributes
related to questions are type, marks and complexity. All these attributes are efficiently used during Question Paper
Generation. The paper generator selects a question according to the pattern and complexity. This engine also introduces
marking systems wherein any selected question is marked so that it might not be selected again. This prevents repetition
of questions in subsequent papers. Finally, generated papers are stored as pdfs.

Copyright to IJARSCT DOI: 10.48175/568 16

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.53 Volume 4, Issue 1, April 2024

IV. ARCHITECTURE DIAGRAM

First the authorized user will register, after the successful registration the user will be provided by the login-Id and
password. After logging-In the user can input the text file either by manually
entering it or by providing the path of the file. After the input is provided, the questions will be
generated accordingly. Questions can be reviewed and manipulated by the user. A question paper can be generated by
selecting desired questions generated by the software. After the generation of question paper, there will be two options:
first it will be emailed to the authorized user or admin and the secondly it will be downloaded as a file.

Algorithm
1. Start
2. Initialization
3. Dataset creation & allocation of marks according to blooms level
4. Preprocess
5. Data Training Process
6. Saving the model for further utilization
7. User login
8. User will upload the CSV of questions
9. Processing the CSV file in system IJIRMPS230356 Website: www.ijirmps.org Email: [email protected] 4
10. Predicting the marks
11. Generating the question paper
12. END

ALGORITHM/TECHNOLOGY
Pre-Processing:
Text pre-processing is a method in natural language processing to prepare the computer understand the structure or
content of the text. It will allow us to prepare the text more precise and easy to use for later process. Text preprocessing
associate processes such as stopwords removal, stemming, lemmatization and POS tagging. In this work, stopwords
removal is applied to the question in order to make the text more readable for later process. Following this, each and
every word will then be tagged using a tagger. In this research, NLTK tagger (Bird et al., 2009) is used to tag the exam
questions. To emphasize the tagging process, consider the following sentence: “Outline how class ArrayList could be
implemented usingan array.”, The tagged output is: Outline/VB how/WRB class/NN ArrayList/NN could/MD be/VB
implemented/VBN using/VBG an/DT array/NN./. The tagger will assist to identify important nouns and verbs, which
may be important in determining the question’s category. In addition, the sentence pattern may help in the correct
recognition of the question’s category. After tagging, some rules will be applies according to question's structure.

Copyright to IJARSCT DOI: 10.48175/568 17

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.53 Volume 4, Issue 1, April 2024

Rules Development:
Through this research, a rule-based approach is accepted in determining the category of an examination question based
on the Bloom’s taxonomy. The rules are developed from a training set which consists of 70 examination questions in
the various subjects. There are two conditionswhere the rules will be applied:
• The rules will distinguish the suitable keywordfor each question depending on its category.
• (elp to choose the appropriate category if the keyword shares more than one category. For example, Abstract may fall
under Comprehension or Synthesis category. After analysing all the questions in the training set, the questions’ patterns
show that most of them start with a verb. (however, only some of it begins with Wh-pronoun, a determiner, preposition
or subordinating conjunction, anoun and an adverb. Before rules can be applied, specific patterns should be identified
from the questions item. The following will determine how the pattern and rules are developed after POS tagging is
applied. Question: Write down the output of the following program:
Question with tag: Write/VB down/RB output/VB following/JJ program/NN :/:
Pattern: /VB (1st word), /VB (3rd word) Each verb in the question will be captured. The verb 'Write' appears as its
keyword. Based on Bloom’s Taxonomy, Write can be categorised into two categories: Knowledge and Synthesis.
FOR each sentence, read into an array. Split into words. )IF pattern is found
)If the keyword “Write” is found
)IF found:
Apply Rule1: Assign weight toKnowledge
Apply Rule2 : Assign weight toSynthesis Choose the greater value or positive value Assign questioncategory Store in
database
FOR EACH(_match in pattern :
print join (keyword, category,question)
Based on the algorithm, the question can be applied to twodifferent rules i.e. Rule 1 and Rule 2. Rule 1 states that the
questions fall under the ‘Knowledge’ category meanwhile Rule 2 states that it can be categorized under the ‘Synthesis’
category. This raises a conflict as to which category the question should fall into. When this situation occurs, there is a
need to introduce ‘category weighting’ to altering all subsequent blocks.
Randomization Algorithm
For N questions available in databaseStep 1: Create a List ‘L’ of N elements
Step 2: Generate a random number ‘n’ such that 1<=nStep 3: )f n ∈ L Go to Step 2
else Store n in the List L
Step 4: Select a question from database corresponding ton, whose flag==true
Step 5: For the question, set flag=false

V. CONCLUSION
Bloom’s Taxonomy is a classification of learning objectives within education that educa- tors set for students. To
automate the process of categorizing examination question according to Bloom’s Taxonomy based on its cognitive
levels. The formation of rules may improve the accuracy of the result .In this system smart model for question paper
generation will be im- plemented as real time application. The proposed work describe a smart system that progress
from the traditional method of paper generation to an smart process by providing the control access to the resources.

REFERENCES
[1]. kithsiri jayakodi,madhushi bandara,Indika pereca.An automatic classifier for exam qua- tions in engineering
:A process for Blooms taxonomy. IEEE international conference on Teching,Assessment and learning for
Engineering(TALE).PAGE:195-202,2015.
[2]. Swart,Author James. Evaluation of final examination papers in engineering : A case study using Blooms
Taxonomy . Education, IEEE Transaction on 53,no.2,p.257- 264,2010.
[3]. Lister, R., and Leaney, ”Introductory programming,criterian-referencing , and Bloom.proceedings of the 34th
SIGCSE technical symposium on computer science education”,Reno,Nevada,USA,SCM Press,2003.

Copyright to IJARSCT DOI: 10.48175/568 18

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.53 Volume 4, Issue 1, April 2024

[4]. Yusof,Norazah, Chai Jing Hui.Determination of Blooms cognitive level of question items using artificial
neural network. In intelligent system design and application (ISDA),2010 10th International conference
on,pp. 866-870,2010.
[5]. Ibrahim EldesokyFattoh ,”Automatic Multiple Choice Question Generation System for Semantic Attributes
Using String Similarity Measures in journal Computer Engineer- ing and Intelligent Systems” www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol.5, No.8,2014
[6]. Ahmed EzzAwad and Mohamed Yehia Dahab,”Automatic Generation of Question Bank Based on Pre-
defined Templates”,in International Journal of Innovations Ad- vancement in Computer Science IJIACS
ISSN 2347 8616 Volume 3, Issue 1 April
[7]. Ms.Deepshree S. Vibhandik, Prof. Mrs.Rucha C. Samant,”An Overview of Automatic Question Generation
Systems”,in International Journal of Science, Engineering and Technology Research (IJSETR), Volume 3,
Issue 10, October 2014 .
[8]. Bloom, Benjamin S.Longman,” Taxonomy of educational objectives: The classification of education goal
.Cognitive domain. Handbook”,1956.
[9]. Ibrahim EldesokyFattoh,”Semantic Based Automatic Question Generation using Artifi- cial Immune
System”,in Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN
2222- 2863 (Online) Vol.5,No.8,2014
[10]. AnnamaneniNarendra, Manish Agarwal and Rakshit shah LTRC,”Automatic Cloze- Questions Generation”,
IIIT-Hyderabad, India,Sept 2013.
[11]. Yvonne SKALBAN , Le An HA , Lucia SPECIA , Ruslan MITKOV,”Automatic ques- tion generation in
multimedia-based learning”,Dec-2012.
[12]. Manish Agarwal, Rakshit Shah and PrashanthMannem,”Automatic Question Genera- tion using Discourse
Cues”,June-2011.
[13]. MingLiuandRafaelA.Calvo,”G-Asks:AnIntelligentAutomaticQuestiongeneration System for Academic
WritingSupport”,2012.
[14]. Andreas Papasalouros, KonstantinosKanaris, KonstantinosKotis,”Automatic genera- tion of multiple choice
questions from domainOntologies”,July-2008.

Copyright to IJARSCT DOI: 10.48175/568 19

www.ijarsct.co.in