E-Learning Recommender System for Teachers using Opinion Mining Essay Example
E-Learning Recommender System for Teachers using Opinion Mining Essay Example

E-Learning Recommender System for Teachers using Opinion Mining Essay Example

Available Only on StudyHippo
  • Pages: 8 (2043 words)
  • Published: August 6, 2017
  • Type: Paper
View Entire Sample
Text preview

E-Learning Recommender System for Teachers using Opinion Mining

The popularity of e-learning as an alternative to traditional classroom learning has increased significantly in recent times. It has transformed education by eliminating geographical limitations and providing access to students with internet connectivity. However, the effectiveness of e-learning websites cannot be guaranteed solely based on their existence. Usually, students need to search multiple websites to gather information on various subjects because no single website offers comprehensive content on a specific topic. Therefore, it is crucial to analyze students' reviews of website content in order to compile the best resources into a centralized platform.

In this paper, we present Angstrom3, a novel e-learning recommender system that utilizes sentiment analysis to provide personalized recommendations. The system suggests instructors who have uploaded tutorials on the website to modify specific difficult aspects of the topic cont

...

ent based on learners' sentiments. Instead of focusing on the entire subject, Angstrom3 employs sentiment mining to analyze individual aspects. This approach allows for the gradual accumulation of the best content about the topic in one convenient location.

Keywords: Opinion mining, Recommender System, User reviews.

Feature Extraction.

  1. Introduction

The rise of cyberspace and associated technologies has led to the widespread use of e-learning by students and learners. The primary objective of the e-learning system is to enhance knowledge delivery in a more efficient manner. It is essential to take into account the learner's perspective on the content topic in order to improve understanding for the majority.

The paper discusses the use of a recommender system called A3, which gathers scholars' opinions on a specific topic. A3 utilizes sentiment mining and feature extraction to identify the aspect of the subject that is difficult fo

View entire sample
Join StudyHippo to see entire essay

learners to understand. The system then suggests instructors to replace this challenging part with a new one, and collects feedback from learners on the modified segment.

The e-learning website will include the top e-material related to a specific subject, making it easier for students to understand. The article is organized into sections: Section 2 discusses e-learning, Section 3 introduces sentiment mining, Section 4 provides an overview of feature extraction, Section 5 presents the functioning of our e-learning recommender system (A^3), Section 6 discusses execution, and Section 7 concludes the article.

  1. E-Learning

In the 1960s, psychologists Patrick Suppes and Richard C., professors at Stanford University, conducted research on psychological science.

Atkinson initially implemented computer-based e-learning to educate elementary school students in math and reading [1]. As IT services have expanded, e-learning has become accessible worldwide. The primary benefit of e-learning is its flexibility, allowing students to study at their convenience from any location [2]. Furthermore, it provides an engaging and user-friendly platform that incorporates various multimedia elements such as text, images, animation, and sound to facilitate effective learning [3].

E-learning is beneficial for individuals with physical disabilities [4]. To ensure the success of an e-learning system, it must be adaptable to the needs of its users (students and instructors) [5], user-friendly, and interactive [6].

  1. Opinion Mining

Opinion mining, also known as data mining, examines individual opinions and sentiments [7, 8, 9, 10, 11]. Our recommender system A3 analyzes students' evaluations on specific subjects.

Using characteristic based sentiment analysis, this approach aims to identify a learner's sentiment towards various topics within

a subject. The system accomplishes this by pinpointing characteristics and sentiment words. By doing so, the system can determine whether expressed sentiments are positive or negative [12].

  1. Feature Extraction

In sentences used for reviewing purposes, subject characteristics usually comprise of nouns or noun phrases, whereas adjectives function as sentiment words.

A3 utilizes the Stanford part-of-speech tagger to parse sentences and generate word part-of-speech tags. The tagged sentences are then stored in a database and employed for extracting features of scholarly reviews related to the subject.

The e-learning tutorial for the Data Structure topic, specifically on Link List, examines various features to determine good and bad characteristics. These features encompass the subject's account, algorithmic portion, programming portion, etc. According to a user review, the algorithm in the Link List tutorial is challenging to comprehend.

Example 1. The Stanford part-of-speech tagger analyzes the sentence and labels each word accordingly as shown in Figure 1 when applied to Example 1.

End product
tally: Loading default properties from trained tagger taggers/left3words-wsj-0-18.tagger Reading POS tagger model from taggers/left3words-wsj-0-18.tagger...done [0.6 sec].Input: Algorithm is difficult to understand. End product of Example 1: Algorithm/NNP is/VBZ hard/JJ to/TO understand/VB./. BUILD SUCCESSFUL (total time: 1 second)

Figure 1.The output of the Stanford part-of-speech tagger that utilizes the standard Penn Treebank tag set can be seen here.
Table 1 provides further details regarding this specific tag set.

Table 1. Penn.The Treebank Tag-set in the

shows various tags and their descriptions. The tags include Milliliter, PRP $, Cadmium, Rubidium, DT, RBR, Ex-husband, FW, Inch, JJ, JJR, JJS, Liter, Mendelevium, NN, NNS, NNP, NNPS, PDT, Polonium, WP, WP $, PRP, and WRB. They are used to classify parts of speech and other linguistic

elements in a sentence. In illustration 1, the sentence demonstrates that the characteristic is 'Algorithm' and the adjectival is 'hard'.

  1. Working of A
    3
    Recommender System

To utilize our recommender system (A3), any e-learning website must ensure that scholars can review tutorials uploaded by instructors. Typically, subject tutorials are divided into subtopics, and it is evident that certain subtopics are well-written and easily understandable for scholars. However, there are also subtopics that are more difficult or not easily understood by many scholars.

For an e-learning website used by thousands of scholars, it can be challenging for instructors to read and make changes to the subject content based on the numerous reviews received. A 3 system is proposed to help instructors identify areas that need improvement in tutorials, making it easier for scholars to understand. The 3 system works as follows:

  1. All reviews about each tutorial are collected and stored in the COLLECTOR, organized by subject in a database.
  2. Stored reviews are then passed to the CLASSIFIER, which categorizes each review as positive, negative, or neutral using SentiWordNet.
  3. For each subject, the NEGATIVE block collects all negatively classified reviews along with their timestamps.
  4. Once the number of reviews for a subject exceeds 10, the COMPARATOR block calculates the total number of reviews (TR) and the total number of negative reviews (NR) for that subject from the NEGATIVE block.The text describes the process flow of the Angstrom 3 recommendation system, which is illustrated in Figure 1. When the TR value is between 1.0 and 2.0, negative reappraisals are sent

to the TOPIC EXTRACTOR block. If the consequence value is greater than 2.0, it is assumed that the tutorial is good and no changes are needed for learners. The FEATURE EXTRACTOR block extracts characteristics from negative reappraisals using Stanford part-of-speech tagger and stores them in the database for the RECOMMENDER block. The RECOMMENDER block generates recommendations for instructors based on the uploaded tutorial, providing subtopics that require improvement from a learner's perspective. The recommended subtopics are then uploaded by the teacher with better illustrations, starting a new cycle of Angstrom 3. The recommendation cycle ends based on the result from the COMPARATOR block. Figure 2 represents the Angstrom 3 recommendation system.The text above describes the practical execution of the Angstrom 3 recommendation system. The instructor will upload the tutorial subject on the e-learning website, which uses the A 3 recommender system. This uploaded acquisition material will be visible on a webpage.

Learners can compose their reappraisals about the tutorial in the same web page as shown in Figure 2.


Linked List

Linked list Theory

It is the aggregation of informations that are connected.

Linked lists are a manner to hive away informations with constructions so that the coder can automatically make a new topographic point to hive away informations whenever necessary.

Algorithm

Step1: struct node /*Create a node */ { int informations ; struct node *next ; } ; Step2: struct node *p ; Step3: p= ( struct node * ) malloc ( sizeof ( struct node ) ) ; Step4: ( *p ) .data=sssssss ; ( *p ) .next=null ; Step5: usage this node in plan.

Your Reappraisal:


Reappraisals:

1.The algorithm is non cleared. 2.It is nice tutorial. 3.The algorithm is absurd. 4.The algorithm is given without clear stairss. 5.Theory is clear.

Figure 2. Web page of e-learning web site. A
3 will hive away information about all the instructors that are uploading larning stuff on to the e-learning web site in
teacher information tabular array shown in table 2.


Table 2.

Teacher Information table.


E mailElectronic mail
< /table>

Table 3. contains information about all the subjects of the topic, known as Subject Topic table.

Table 3.
Capable Topic table.


Field Name

Description
TiD Teacher unique ID
SiD Capable Teacher Dealing with
< td >

Field Name
< td >< strong >Description

< tr >

SID< td >Capable unique Id< tr >< t d>T opicIDp ic uni que Id/t r>

T a ble4 .

The tutorial information for each subject is stored in a table called "Tutorial tabular array" (Table 4). This table includes the fields Field Name and Description. The

Field Name column lists the names of the fields, while the Description column provides a description for each field. The fields within this table are TopicID, Content, and Date.

Another table, called "Review tabular array" (Table 5), stores all reviews for a subject along with their classification done by the CLASSIFIER. This table has similar fields as the previous one, including TopicID, ReviewId, ReviewContent, Reappraisal date, and ReviewClass.

The "Feature Extractor" table (Table 6) contains subtopics extracted from negatively classified reviews. It consists of two fields: TopicID and SubtopicName.

Finally, the "Recommender" table (Table 7) is used to generate recommendations for the instructor.The table contains five fields, namely "TopicID" which represents the unique ID of the topic, "SubtopicName" indicating the name of the subtopic that requires improvement, "TiD" representing teacher's Idaho alone, "Date" showing the date when Recommendation Generation occurred, and "RFlag" containing additional information related to recommendation. By utilizing tables 2 to 7, a A3 recommender system will generate recommendations for instructors regarding subjects they have uploaded on an e-learning website.


  1. Decision

An A

3

recommender system employs sentiment mining to understand the problem faced by scholars while studying a specific subject. The system identifies the specific subtopic where the scholar is encountering difficulties. It then identifies the instructor responsible for that subtopic and generates recommendations for them. These recommendations consist of the subtopics that need further explanation for the scholars.

The text mentions Naidu, S. (2005) from the Measuring Distance Association of Australia and Schenk, P., who participated in a conference held on 9-11.

Rogers, A. G. Berg

(Eds.), Naidu, S., Oliver, M., and Koronios (1999).

Group, Inc. : Hershey PA. 2. B.

Khan, a book titled Web-based Instruction, was published by Education Technology Publication in 1997. The authors of the book are R. Clark and A. R.

Mayer, E-learning and the Science of Instruction: A book that provides proven guidelines for consumers and interior decorators of multimedia acquisition, published by Jossey-Bass in 2003 (chapter 4).

Tseng (2004) conducted a study on the web usability of Chinese online learning websites. The findings were published in the Journal of National Taipei Teachers College Education, volume 17 (1), pages 271-298.

L. C. Seng and T. T. Hok presented a paper titled "Humanizing E-learning" at the 2003 International Conference on Cyber Universes in Singapore [6].

H. Giroire, F. Le Calvez, G. Tisseau ; "Benefits of knowledge-based synergistic acquisition environments: A instance in combinatorics" ; Proceedings of the Sixth International Conference on Advanced Learning Technologies, 2006.

pp:285-289. 7. G. Vinodhini, R.M. Chandrasekaran, Sentiment analysis and sentiment excavation: a study, Int.

J. Adv. Res. Comput.

Sci. Software Eng. 2 (6) (2012). 8. M.

Chau, J. Xu conducted a survey on online hatred groups, specifically focusing on mining communities and the relationships within web logs in the article "Mining communities and their relationships in web logs: a survey of online hatred groups" published in the International Journal of Human-Computer Studies (2007, vol. 65, issue 1).

9. A. Reyes, P. Rosso, D.

Buscaldi, From wit acknowledgment to irony sensing: the figurative linguistic communication of societal media, Data Knowl. Eng. 74 (2012) 1–12. 10.

F. Salvetti, S. Lewis, C. Reichenbach conducted a study on the automatic sentiment mutual opposition categorization of film reappraisals in the Colorado Research in Linguistics journal in 2004.

11. T.S. Raghu

and H. Chen discuss the advancements in information sharing, information excavation, and collaboration systems for the fatherland security in their paper "Cyberinfrastructure for fatherland security", published in the Decis. Support Syst. journal.

43 (4) (2007). 12. Padmapani P. Tribhuvan, S.G. Bhirud, Amrapali P.

The article titled "A Peer Review of Feature Based Opinion Mining and Summarization" by Tribhuvan was published in the International Journal of Computer Science and Information Technologies (IJCSIT) in 2014.

Get an explanation on any task
Get unstuck with the help of our AI assistant in seconds
New