proposalbeamer.pdf

7/27/2019 proposalbeamer.pdf

1/49

IntroductionReview of Related Literature

Methodology

TECS: Test Essay Checking Software

Marlon Fernando, Efren Ver Sia, Cheryleigh Anne Verano andProspero C. Naval, Jr.

Computer Vision & Machine Intelligence GroupDepartment of Computer Science

College of Engineering

University of the Philippines-Diliman

July 8, 2013

1/49


2/49


Methodology

ProblemsStatement of the HypothesisObjectives

Outline

1 IntroductionProblemsStatement of the HypothesisObjectives

2 Review of Related LiteratureExisting SystemsConcept Indexing

3 MethodologyProcedures and AlgorithmsEquipment SpecicationsEntity Relationship DiagramData Collection and Analysis

2/49


3/49


Methodology


Problems

3/49


4/49


Methodology


Problems

Essay-type examination is a useful evaluation tool to fullycapture a students knowledge on a certain topic.

4/49


5/49


Methodology


Problems

Essay-type examination is a useful evaluation tool to fullycapture a students knowledge on a certain topic.Evaluating essays can consume a huge amount of time andeffort.

5/49


6/49


Methodology


Problems

Essay-type examination is a useful evaluation tool to fullycapture a students knowledge on a certain topic.Evaluating essays can consume a huge amount of time andeffort.Teachers are challenged to be consistent on grading essays toensure no bias and same standards are applied in all essays.

6/49


7/49


Methodology


Problems

Essay-type examination is a useful evaluation tool to fullycapture a students knowledge on a certain topic.Evaluating essays can consume a huge amount of time andeffort.Teachers are challenged to be consistent on grading essays toensure no bias and same standards are applied in all essays.Scores of human raters can be affected by certain factors likefatigue, mood, and distractions .

7/49


8/49


Methodology


Problems

Essay-type examination is a useful evaluation tool to fullycapture a students knowledge on a certain topic.Evaluating essays can consume a huge amount of time andeffort.Teachers are challenged to be consistent on grading essays toensure no bias and same standards are applied in all essays.Scores of human raters can be affected by certain factors like

fatigue, mood, and distractions .Large-scale assessments are usually limited to multiple-choicequestion type.

8/49


9/49


Methodology


Statement of the Hypothesis

TECS will be able to produce essay score and feedback that isalmost as good as human evaluation results.

9/49


10/49


Methodology


Objectives

10/49


11/49


Methodology


Objectives

Provide a fast and more accurate essay scoring by eliminatinghuman rater fatigue and inter-rater reliability issue

11/49

I d i P bl


12/49


Methodology


Objectives


Highly reduce cost of effort and time consumption for gradingmultiple essays

12/49

I t d ti P bl


13/49


Methodology


Objectives


Highly reduce cost of effort and time consumption for gradingmultiple essaysAble to give an instant feedback for an essay

13/49

Introduction Problems


14/49


Methodology


Objectives


Highly reduce cost of effort and time consumption for gradingmultiple essaysAble to give an instant feedback for an essayProvide a fast and memory-efficient algorithm for essay

grading

14/49



15/49


Methodology


Objectives



gradingIntroduce a breakthrough in education

15/49



16/49


Methodology


Objectives



gradingIntroduce a breakthrough in education

16/49

Introduction


17/49


MethodologyConcept Indexing

Outline




17/49

Introduction


18/49

Review of Related LiteratureMethodology

Concept Indexing

18/49

Introduction


19/49


Concept Indexing

19/49

Introduction


20/49


Concept Indexing

Concept Indexing

20/49

Introductionf l d d


21/49


Concept Indexing

Concept Indexing

Starts with tokenization of words in an essay and thegeneration of Term-by-Document Matrices

21/49

IntroductionR i f R l t d Lit t C t I d i


22/49


Concept Indexing

Concept Indexing

Starts with tokenization of words in an essay and thegeneration of Term-by-Document MatricesInstead of SVD, it uses Concept Decomposition (CD).

22/49

IntroductionReview of Related Literature Concept Indexing


23/49


Concept Indexing

Concept Indexing

Starts with tokenization of words in an essay and thegeneration of Term-by-Document MatricesInstead of SVD, it uses Concept Decomposition (CD).K-means Clustering VS. Fuzzy C-means Clustering

23/49



24/49


Concept Indexing

Concept Indexing

Starts with tokenization of words in an essay and thegeneration of Term-by-Document MatricesInstead of SVD, it uses Concept Decomposition (CD).K-means Clustering VS. Fuzzy C-means ClusteringIn experiments that were conducted, both versions of CIoutperform LSI in Exact Agreement Accuracy andPearsons Product-Moment Correlation Coefficient

24/49



25/49


Concept Indexing

Concept Indexing

Starts with tokenization of words in an essay and thegeneration of Term-by-Document MatricesInstead of SVD, it uses Concept Decomposition (CD).K-means Clustering VS. Fuzzy C-means ClusteringIn experiments that were conducted, both versions of CIoutperform LSI in Exact Agreement Accuracy andPearsons Product-Moment Correlation Coefficient

In particular, CI with Fuzzy C-means achieves better resultsthan CI with K-means in essay content evaluation, and CIwith Fuzzy C-means is less sensitive to the preprocessingtechniques such as stemming and stopwords removal andeventually, to noise.

25/49


Procedures and AlgorithmsEquipment SpecicationsE i R l i hi Di


26/49

Review of Related LiteratureMethodology Entity Relationship DiagramData Collection and Analysis

Outline




26/49


Procedures and AlgorithmsEquipment SpecicationsE tit R l ti hi Di


27/49

Methodology Entity Relationship DiagramData Collection and Analysis

Procedures and Algorithms

27/49


Procedures and AlgorithmsEquipment SpecicationsEntity Relationship Diagram


28/49



Spelling Grammar Checker: GNU Aspell/LanguageTool

28/49




29/49



Spelling Grammar Checker: GNU Aspell/LanguageToolContent Analysis: Concept Indexing

29/49




30/49



Spelling Grammar Checker: GNU Aspell/LanguageToolContent Analysis: Concept IndexingAgile software development

30/49


M h d l



31/49


Equipment Specications

31/49


M th d l g



32/49

Methodology y p gData Collection and Analysis


a web hosting service subscription (approx. P200.00 per

month) or a dedicated server with 2-4 GB RAM (approx.P6,000.00 per month) that runs Apache, Linux andPostgreSQL

32/49


Methodology



33/49

Methodology y p gData Collection and Analysis



month) or a dedicated server with 2-4 GB RAM (approx.P6,000.00 per month) that runs Apache, Linux andPostgreSQLLanguage: PHP(backend) and HTML/JS/CSS (frontend)

33/49


Methodology



34/49

Methodology Data Collection and Analysis



month) or a dedicated server with 2-4 GB RAM (approx.P6,000.00 per month) that runs Apache, Linux andPostgreSQLLanguage: PHP(backend) and HTML/JS/CSS (frontend)Framewok: Yii

34/49


Methodology


ll d l


35/49


Entity Relationship Diagram

35/49


Methodology

Procedures and AlgorithmsEquipment SpecicationsEntity Relationship DiagramD C ll i d A l i


36/49


Entities

36/49


Methodology

Procedures and AlgorithmsEquipment SpecicationsEntity Relationship DiagramData Collection and Anal sis


37/49

gy Data Collection and Analysis

Entities

USERS:1 Attributes: rst name, last name, middle name, and username2 Roles: 0 for admin, 1 for school administrator, 2 teacher,

and 3 for the students3 The slug eld is composed of characters that will be sent to

his/her mail for verication of accounts, and veried eld is todetermine if the users already veried his/her account.

37/49


Methodology

Procedures and AlgorithmsEquipment SpecicationsEntity Relationship DiagramData Collection and Analysis


38/49

gy Data Collection and Analysis

Entities




BASIC INFO: consists of users mobile number and his/heraddress: the street, municipality and province.

38/49


Methodology



39/49

Data Collection and Analysis

Entities




BASIC INFO: consists of users mobile number and his/heraddress: the street, municipality and province.

CLASS INFO:1 class name and class key separated from the class list of a

teacher2 The user id eld is for the id of the teacher who created the

class

39/49


Methodology



40/49


Entities

40/49


Methodology



41/49


Entities

CLASS:1 This table is for the students who will enroll for the class.2 It stores the user id of the students, and the class id of the

class that was made by the teacher.

41/49


Methodology



42/49

y

Entities


class that was made by the teacher.EXAMS INFO: This is where the essays of the students will be

saved, and the date when the student passed his/her essays.

42/49


Methodology



43/49

Entities



saved, and the date when the student passed his/her essays.RESULT:1 It consists of the nal score given to the essay that has been

passed by the students.2 It also has the feedback of the teacher if he/she wants to put

comments in the essay of a particular student.3 And the exam id is from exam info where the body of essay

resides.

43/49


Methodology



44/49

Entities



saved, and the date when the student passed his/her essays.RESULT:1 It consists of the nal score given to the essay that has been

passed by the students.2 It also has the feedback of the teacher if he/she wants to put

comments in the essay of a particular student.3 And the exam id is from exam info where the body of essay

resides.EXAMS: This stores the user id of the student who passed anessay, an exam id where the body of essay resides, and theclass id where the student is enrolled.

44/49


Methodology



45/49


45/49


Methodology



46/49


The proponents will ask at least four teachers in elementary orin high school for their evaluation on at least 50 studentessays each.

46/49


Methodology



47/49


The proponents will ask at least four teachers in elementary orin high school for their evaluation on at least 50 studentessays each.We shall have two sets of essays: one will consist of technicalor objective essay and the other set will consist of creative orsubjective essays.

47/49


Methodology



48/49


The proponents will ask at least four teachers in elementary orin high school for their evaluation on at least 50 studentessays each.We shall have two sets of essays: one will consist of technicalor objective essay and the other set will consist of creative orsubjective essays.The teachers evaluation results will then be compared to themachines and will be subjected to Pearsons r for the analysis

of their correlation.

48/49


Methodology



49/49


The proponents will ask at least four teachers in elementary orin high school for their evaluation on at least 50 studentessays each.We shall have two sets of essays: one will consist of technicalor objective essay and the other set will consist of creative orsubjective essays.The teachers evaluation results will then be compared to themachines and will be subjected to Pearsons r for the analysis

of their correlation.For additional training and testing data set, the proponentswill also use essays fromKaggle.

49/49

proposalbeamer.pdf

Documents