PREDICTION MODEL FOR INTERNSHIP DEPLOYMENT USING …...Internship programming is acquiri. ng practical skills bef. ore they can accept in different companies [10]. Intern Students

PREDICTION MODEL FOR INTERNSHIP DEPLOYMENT

USING DECISION TREE Anna Liza A. Ramos, Daniel B. Quijano, Karizza Reigne E. Infante, Jomar M. Aviles, Wilmar D. Olitoquit,

Christian Pagalanan 1Saint Michael’s College of Laguna, Biñan Laguna, Philippines

Email: [email protected]

ABSTRACT

This paper aims to predict the performance of Computer Science students based on their General Point Average (GPA) in Eng-lish, Mathematics and Programming courses as basis for internship deployment. The study applied decision tree algorithm to classify the student performance with prediction result of 90.91% for Competent, 66.67% for Capable and 100% for Challenged. This result will provide the institution to assess the preparedness of the students in the workplace. Keywords : Internship, Decision Tree, Students performance (keywords)

1 INTRODUCTION

ata may be a kind of method to remodel information into helpful data and is significant for any tiny or additional significant business for creating strategic selections. The

advancement of technology makes information today is ana-lyzed and understood supported the models. With the ad-vancement of technology, machine learning was developed to automatize the method in predicting the outcomes of the info in step with the algorithmic rule used. what is more, there are tools accustomed to analyzing information like RapidMiner, Knime OpenRefine, Gephi, and NodeXL [13] that provides information table, chart, images, vector-like and the like. Data Analytics is the technique of separating helpful insights from data mistreatment algorithms and techniques that may discover hidden patterns. information analytics involves the employment of analytics techniques like machine learning, data processing, tongue process, and statistics it additionally involves the employment of advanced techniques and tools of analytics on the info obtained from different sources in nu-merous sizes. And different analytics like business refers to the applications, skills, technologies, and practices for contin-uous reiterative exploration to understand past the perfor-mance to produce unjust insights. And focuses on developing and understanding the performance supported information and applied mathematics ways [5]. Moreover, information analytics comes in to deal with the massive data, convenience for the user and lessen the overwhelming to calculate this in-formation.

1.1 BACKGROUND OF THE STUDY

There are several studies conducted to predict student per-formance using Decision Tree such as a study to identity stu-dents in advance who fails or not for the professors to enhance their teaching skills and to give more time and effort for the failing students with the success rate of 60.46% in out of 346 instances 209 instances are correctly classified and accurate in a small data (R. R. Kabra, et.al 2014) [17]. A study on the per-

formance in Internet Technology and Programming course which includes BFTree and CART C4.5 Algorithm with a clas-sification rate of 67.07% (Abdulsalam Sulaiman Olaniyi, at.al 2017) [19]. A study to measure student examination perfor-mance using decision tree to determine student’s capability (Dr. Anjali B Raut, et.al 2017) [2]. A study to predict the final GPA to determine if the student get accepted in colleges with the success rate of 87% (Mashael A. Al-Barrak, et.al 2016) [12]. A study on the academic performance using the four algo-rithms J48, NBtree, Reptree and Simple cart however, among the algorithms J48 decision tree algorithm is best suitable algo-rithm with the accuracy rate of 80.15% (Mrinal Pandey, et.al 2014) [14]. These studies provide great ideas for the study to utilize the Decision Tree in predication academic performance of students who will qualify for On-the-Job training using General Point Average of their Programming, Mathematics and Communication courses in order to ensure that the stu-dents are competitive and prepared to execute their acquired skills in their field of profession.

One of the purposes of this proposed study is that the

developed system analyzes the academic performance of the

student to use as a basis if a student is capable to be deployed

in an internship, it is an opportunity offered by an employer to

potential employees which are called as interns. This is to

work at a fixed period and usually take about 200 to 600 hours

or may last between one month and up to three months [7].

Most of the interns are undergraduates or students and they

are usually part-time if offered during a university semester

and full-time if offered during the vacation. The reasons for

taking up an internship are this helps the students to gain ex-

perience in the real world, to have a better understanding on

your role as an intern, tasks and in the industry, to get the op-

portunity to learn and watch, to acquire the ability to practice

new things, this also helps you to build your confidence and

D

IJOART

International Journal of Advancements in Research & Technology, Volume 8, Issue 4, April-2019 ISSN 2278-7763 15

IJOART Copyright © 2019 SciResPub.

to practice your communication as well and to help you

choose on what path that you are going to take if you are so

unsure of what to do in the near future [6].

1.2 RESEARCH OBJECTIVE

To utilize the academic grade point as basis to predict student per-

formance for their internship

To implement a prediction model using Decision Tree.

To build a prototype model that will provide visual information

1.3 Theoretical Framework

1.3.1 On-the-job Training

Communication skills will contribute to our interpersonal or people skills. also, the tone of voice, body language and other non-verbal cues send messages [11]. To communicate well and effectively share the information in manner to easily under-stand [3]. OJT is the training ground of the students to develop their communication skills and challenges in a big world or working industry to enhance their skills to communicate well in others [1]. And to training them to perform in real world. Communication skills are the main requirements in the com-panies [8]. Before they can accept the students in On the Job Training.

1.3.2 Computer Science Competencies

Internship programming is acquiring practical skills before

they can accept in different companies [10]. Intern Students

can have a critical thinking skills, basic skills, personal quali-

ties and competencies after they finished their OJT [9]. Positive

results may affect the student that is deployed as intern in the

workplace specifically as their confidence in communicating

will boost and working with the others is not a problem and

focusing on their career path that they are taking [4].

1.3.3 Decision Tree

A decision tree is a support device that uses a tree-like chart or model of decision and their conceivable results, including chance occasion results, expenses, and utility. It is one ap-proach to show a calculation that lone contains restrictive con-trol articulations. It is commonly used in operations research, specifically decision analysis and to help identify a strategy that is more likely a goal, but a favorite tool in machine learn-ing. It is also a hierarchical tree structure that used to charac-terize classes because of an arrangement of questions about the qualities of the class. Some studies are (EMIL LUNDKVIST, et.al 2014) To display and interact with the ex-isting database with price information at the company. This study is to work fast in real time to analyze the products in different classes. It is fast enough to operate in real time and performs well. They used a decision tree to divide the rest of the products in the database into classes then and each prod-uct within the class can be predicted. (Adewale Opeoluwa

Ogunde, et.al 2017) The study focuses on developing a system that showed the hidden relationships between the crime-related data, in form of decision trees to detect and improve the favorable and high rates of success to imitate the crime in university. Before they came up to this research, they collect data both crimes and criminals to develop the test model. The study is very efficient for detecting crimes because it merely analyzes the data retrieved from existing data in the database with an accuracy of 72.7%. (Nazmun Nahar, et.al 2018) To cal-culate the performance of various decision tree techniques and compare their performance. This study is to identify the pa-tient whether he/she has liver disease or not. This study used J48, LMT, Random Forest, Random tree, REPTree, Decision Stump, and Hoeffding Tree but Decision Stump is the highest accuracy rate than other techniques. (Mashael A. Al-Barrak, et. al 2016) To predict students' final GPA based on their grades in previous courses. This study is to identify the student’s GPA too easy to know which colleges are acceptable in their grades. The study used the J48 decision tree classification al-gorithm to easily predict the student’s GPA. This study reached an overall accuracy of 87%.

2. METHODOLOGY

2.1 Data Gathering

The GPA were collected to selected 20 students with their permission. This is reflected in their prospectus consist of 4 Communication subjects, 4 Mathematics and 18 Computer Science Professional Courses. The GPA are classified 1.00 – 1.50 as Competent, 1.75 – 2.25 as Capable, and 2.50 – 3.00 as Challenged. The students have 4 Communication courses, 4 Mathematics courses and 18 Programming courses to be com puted for prediction. Generating the Model

IJOART



The study used RapidMiner to create the Decision Tree Mode utilizing student’s GPA in Programming, Mathematics, and Communication. The model will predict student classification together with their respective prediction result and it provides accuracy result of the classification.

2.2 Classification Model

Figure 1. Decision Tree Model

The study generated a Decision Tree of the student’s classifica-tion by inputting their average grades in Programming, Communication, and Mathematics. An If-Then Else-Then Statement is also provided for programmatic guidance. RapidMiner seem to have put Programming as the root node of the tree. The study found this acceptable since it is the most important factor a company would look for in an intern. It is also noticeable in the generated tree that, if a student gets an average higher than 2.10 in Programming, he/she can already receive his/her overall classification by inputting the Mathe-matics grade. Otherwise, if the Programming average grade is lower than 2.10, he/she would receive his/her overall classifi-cation after Communication grade and Mathematics grade has been inputted. Under this is the Decision Tree Condition Statement.

2.3 Prototype Model

Figure 2. Prototype Model

The prototype model is developed with PHP, HTML, and CSS to create the user interface and XAMPP as the server and Notepad++ as the compiler. The prototype shows the graph-ical representation of average, student grades and over-all academic performance of the students.

3 RESULTS AND DISCUSSION

Table 1: Prediction Result of Student Performance using GPA.

true Com-

petent true

Capable true Chal-

lenged class

precision

pred. Competent

1 10 0 90.91%

pred. Ca-pable

2 0 0 100%

pred. Chal-lenged

0 1 6 85.71%

class recall 90.91% 66.67% 100%

The table shows the predicted result of the Decision Tree to the student datasets that needed Overall Classification. The model has predicted and classified the students, whether they are Competent, Capable or Challenged, with and accuracy of 90%. Fig. 3.1 shows a split in confidence of a prediction in two students. Though the final prediction was correct when manu-ally calculated, RapidMiner regarded this situation as a de-crease in accuracy for the model. Table 2: Split Confidence Prediction

Studnum Prediction (Overall)

Confidence (Capable)

Confidence (Competnt)

Confidence (Challenge)

Student3 Capable 0.500 0.500 0

Student12 Capable 0.500 0.500 0

This shows that there are students where the model wasn’t confident of the classification which create a negative to accu-racy of the classification.

if (Programming >= 2.10) || (Communication >= 2.40)

|| (Mathematics >= 1.90) THEN Competent || (Mathematics < 1.90) THEN Capable

|| (Communication < 2.40) THEN Capable if (Programming < 2.10)

|| (Mathematics >= 1.90) THEN Capable || (Mathematics < 1.90) THEN Challenged

IJOART



4 CONCLUSION AND RECOMMENDATION

This study aims to predict, formulate, construct, and evaluate student Internships accuracy. The developed model is well-suited for the analysis of many objective factors, for students with the qualified knowledge, and skills in Company that they pursue to internships. The experimental results show that the proposed prediction model has good prediction effect and excellent performance. Predicting student Internships will help to make significant improvements in the experience and skills of students and help them to discover the world of job experience and to get in touch with people. For the im-provement of the study.

Add more relative factors to provide accurate prediction.

Apply other classification algorithm to determine the

best performing algorithm for this data.

ACKNOWLEDGMENT

The authors are grateful to the learning applied by the school in order to increase our level of understanding and to the pan-el members who are passionate to guide us to make this study produce valuable information.

REFERENCES

[1] A. Bernardo, A. Landicho, J. M. Laguador (2014) "On-the-Job Training Per-

formance of Students from AB Paralegal Studies for SY 2013-2014" Studies in

Social Sciences and Humanities Vol. 1, No. 4, 122-129

[2] Dr. A. B. Raut & Ms. A. A. Nichat (2017) ‘Students Performance Prediction

Using Decision Tree Technique ‘International Journal of Computational Intel-

ligence Research ISSN 0973-1873 Volume 13, Number 7 (2017), pp. 1735-1741.

[3] E. Bhattacharyya, S.B.M. Nordin, R.B. Salleh (2014) "Internship Students'

Workplace Communication Skills: Workplace Practices and University Prep-

aration" 2014

[4] E. J. C. Valdez, S. S. B. Alcantara, C. A. Pamintuan, J. G. Relos, R. C. Castillo

(2015) "Contributions of On-the-Job Training Program to the Skills, Personal

Qualities and Competencies of Tourism Graduates"

[5] He, Xin James (2014) "Business Intelligence and Big Data Analytics: An Over-

view," Communications of the IIMA: Vol. 14: Iss. 3, Article 1.

[6] M. A. Al-Barrak and M. Al-Razgan. ‘Predicting Students Final GPA Using

Decision Trees: A Case Study’. International Journal of Information and Edu-

cation Technology, Vol. 6, No. 7, July 2016

[7] M. Pandey and V. K. Sharma, PhD. (2014) ‘A Decision Tree Algorithm Per-

taining to the Student Performance Analysis and Prediction’. International

Journal of Computer Applications (0975 – 8887) Volume 61– No.13,

[8] M. Rouse and C. Stedman (2016) “Search Data Management: Data Analytics”

[9] M. Sanon (2017) “4 Reasons Why Data Analytics is Important: Data Analytics

Blog”

[10] R. R. Kabra & R. S. Bichkar (2014) ‘Performance Prediction of Engineering

Students using Decision Trees’. International Journal of Computer Applica-

tions (0975 – 8887) Volume 36– No.11, December 2014

[11] R.T. Cineca (2015) “Introduction to Data Analytics School on Scientific Data

Analytics and Visualization” June 2015

[12] S. O. Abdulsalam, Y. K. Saheed, M. A. Hambali, T. T. Salau-Ibrahim, N. B.

Akinbowale (2017) ‘Student’s Performance Analysis Using Decision Tree Al-

gorithms’ Anale. Seria Informatică. Vol. XV fasc. 1 – 2017.

[13] Z. Osborne, D. Houle, K. Davis, S. Escobedo (2014) "Academic Internship

Program – Students"

[14] R. Casiple (2014) “OJT Program: The most effective form of job training”

IJOART



PREDICTION MODEL FOR INTERNSHIP DEPLOYMENT USING …...Internship programming is acquiri. ng practical skills bef. ore they can accept in different companies [10]. Intern Students

Documents