Article Per Year (5 Year)
p-Index From 2021 - 2026
1.607
P-Index
This Author published in this journals
All Journal IJISTECH (International Journal Of Information System & Technology) Akrab Juara : Jurnal Ilmu-ilmu Sosial IJISTECH
Perangin-Angin, Elvi Sunita
Unknown Affiliation
Biochemistry, Genetics & Molecular Biology Chemistry Computer Science & IT Decision Sciences, Operations Research & Management Electrical & Electronics Engineering Engineering Medicine & Pharmacology Social Sciences

Published : 10 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 4 Documents
Search
Journal : IJISTECH

 1 
Improved Naive Bayes Algorithm with Particle Swarm Optimization to Predict Student Graduation Wati, Embun Fajar; Perangin-Angin, Elvi Sunita; Sari, Anggi Puspita
IJISTECH (International Journal of Information System and Technology) Vol 7, No 6 (2024): The April edition
Publisher : Sekolah Tinggi Ilmu Komputer (STIKOM) Tunas Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30645/ijistech.v7i6.338

Abstract

Timely graduation is very important for educational institutions such as universities, especially for students. Because it can prove that the University and students are able to undergo the learning process theoretically and practically. But many students do not pay attention to graduation, especially those who are already working or married. Therefore, analysis is needed to predict student graduation so that solutions can be found by the University. Data mining was chosen as a method to process data to get new information. The algorithm used in data mining is Naïve Bayes. The research stages include loading data into excel, cleaning empty data, selecting databases related to graduation and taking data from 300 students majoring in Informatics Engineering. The next stage is data transformation by categorizing student data, namely personal data attributes (gender, age, marital status, job status) and academic data (grade). Data testing, application of Naïve Bayes algorithm and accuracy testing were carried out with Rapis Miner software version 10.3.001. The results of data processing with Rapid Miner using the Naïve Bayes algorithm are shown with the Confusion Matrix and ROC Curve. The results of confusion matrix from data processing with Naïve Bayes in the form of accuracy, precision, and recall have the same result of 100%. The percentage of the Confusion Matrix indicates that the model created can classify correctly and accurately. The ROC curve depicted with AUC yields a value of 1, which means that the test showed excellent results
Prediction of Student Graduation using the K-Nearest Neighbors Method Wati, Embun Fajar; Perangin-Angin, Elvi Sunita; Sari, Anggi Puspita
IJISTECH (International Journal of Information System and Technology) Vol 7, No 3 (2023): The October edition
Publisher : Sekolah Tinggi Ilmu Komputer (STIKOM) Tunas Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30645/ijistech.v7i3.318

Abstract

Predictions on the accuracy of student graduation are designed to support study programs in guiding students so that they can graduate on time. The number of student graduations will influence the university's accreditation score. Graduation predictions can provide very useful information in decision-making; therefore, research was conducted on student graduation data. This data will be processed using the K-Nearest Neighbor method. The dataset used consisted of 150 students majoring in informatics engineering. The variables included gender, age, marital status, grade, and job status. The research methodology used in this study consists of 6 stages: Data Collection, Data Selection, Preprocessing, Transformation, Testing, and Evaluation. In the preprocessing or cleaning stage, the data can be fully utilized because all fields have been filled in correctly. Meanwhile, in the transformation stage, the data is categorized as follows: age (young: 19-24, old: 25-50) and grade (large: 3-4, small: 1-2.9). The K-Nearest Neighbor (KNN) method can predict student graduation rates. The KNN method, processed with the RapidMiner 9.9 tool, obtained an average accuracy of 100%. Based on the results of 100% accuracy and an AUC value of 1, it can be concluded that the KNN method is highly accurate in classifying graduation data for the 150 students.
Customer Loyalty Classification with Comparison of Naive Bayes, C4.5, and KNN Methods Wati, Embun Fajar; Perangin-Angin, Elvi Sunita; Indriyani, Luthfi
IJISTECH (International Journal of Information System and Technology) Vol 8, No 3 (2024): The October edition
Publisher : Sekolah Tinggi Ilmu Komputer (STIKOM) Tunas Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30645/ijistech.v8i3.361

Abstract

Customer loyalty is indispensable for the survival of a company. Customer loyalty needs to be maintained in order to return to visit and transact with the Company. Customer data consisting of age, annual income, purchase amount, region, purchase frequency, and loyalty score features can produce new information, namely analyzing customers who have high loyalty. Data processing is carried out using three data mining algorithms, namely Naïve Bayes, C4.5 or Decision Tree, and KNN. The stages carried out in data processing consist of data selection, preprocessing, transformation, and modelling. The customer data used amounted to 238. Modelling is carried out using Rapid Miner Software. Customer loyalty classification can be easily done with the three algorithms, namely Naive Bayes, and C4.5 or Decision Tree, and KNN which is validated by the 10-fold cross-validation method so as to produce the highest percentage of accuracy and the similarity of the accuracy value of the Naive Bayes and C4.5 algorithms, which is 96.67%. In the AUC value, it can be seen that the Naive Bayes algorithm is superior to the C4.5 algorithm or Decision Tree and KNN. The result of the highest AUC value is 0.997, the highest precision percentage is 98.92% achieved by the Naive Bayes algorithm. The result of the highest recall percentage is C4.5 of 100%. The results of the AUC value and accuracy percentage on the three algorithms prove that the performance of the three algorithms is very good.
Comparison of Naive Bayes and C4.5 Methods with Particle Swarm Optimization on Customer Loyalty Classification Wati, Embun Fajar; Perangin-Angin, Elvi Sunita; Indriyani, Luthfi
IJISTECH (International Journal of Information System and Technology) Vol 8, No 6 (2025): The April edition
Publisher : Sekolah Tinggi Ilmu Komputer (STIKOM) Tunas Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30645/ijistech.v8i6.382

Abstract

The Company attaches great importance to customer loyalty for the sustainability of the Company. Loyal customers will buy many times and provide great profits. In this study, the decision tree method or C4.5 and naïve bayes were used with PSO optimization for customer classification which aims to design a strategy in decision-making towards disloyal customers. Some of the stages carried out are data load into MS. Excel, data cleaning from noise, data selection as many as 238 obtained from previous research with several attributes, including, namely age, annual income, purchase amount, region, purchase frequency, and loyalty score, as well as data transformation, namely each attribute is grouped into 2 with their own criteria, data testing by modeling data through Rapidminer, Data evaluation by examining the values of accuracy, precision, recall, and AUC. Both methods have the same accuracy value of 96.67% and the same recall value of 100%. For the precision value, there is a difference of 0.6% and the precision decision tree value is higher than the naïve Bayes which is 96.16%. As for the AUC value, it is higher naïve bayes, which is 0.922 with the difference from the decision tree of 0.059. It can be concluded that the two methods in processing customer loyalty data in this study have the same accuracy, so both methods are equally good.
Co-Authors Agustiena Merdekawati, Agustiena Embun Fajar Wati Luthfi Indriyani Masshitah, Sari Sari, Anggi Puspita