Claim Missing Document
Check
Articles

Found 33 Documents
Search

Klasifikasi Data Mahasiswa Lampau Menggunakan Metode Decision Tree dan Support Vector Machine Kurniawati, Kurniawati; Kusumawati, Ririen; Yaqin, Muhammad Ainul
ILKOMNIKA Vol 6 No 3 (2024): Volume 6, Nomor 3, Desember 2024
Publisher : Lembaga Penelitian dan Pengabdian Masyarakat

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.28926/ilkomnika.v6i3.687

Abstract

Penelitian ini bertujuan untuk mengklasifikasikan data mahasiswa guna memprediksi kelulusan di Universitas Islam Negeri Maulana Malik Ibrahim Malang, dengan fokus pada identifikasi faktor-faktor yang memengaruhi keberhasilan kelulusan dan potensi putus kuliah. Data penelitian mencakup atribut akademik seperti Indeks Prestasi Kumulatif (IPK) dan jalur masuk, yang digunakan untuk mengelompokkan mahasiswa ke dalam kategori lulus atau putus. Metode klasifikasi yang diterapkan adalah Decision Tree (DT) dan Support Vector Machine (SVM). Decision Tree bekerja dengan membangun model berbasis pohon keputusan berdasarkan atribut paling berpengaruh, sedangkan SVM menggunakan hyperplane optimal untuk membedakan kategori. Dataset mahasiswa angkatan 2018 dianalisis menggunakan Python dan library scikit-learn. Hasil menunjukkan bahwa Decision Tree mencapai akurasi sebesar 96,91%, sedikit lebih tinggi dibandingkan SVM yang mencapai 96,62%. Hasil ini mengindikasikan keunggulan Decision Tree dalam memprediksi kategori kelulusan. Penelitian ini berkontribusi pada pemanfaatan atribut akademik sebagai indikator utama dalam klasifikasi data kelulusan serta membandingkan efektivitas dua algoritma pada konteks pendidikan. Dengan temuan ini, universitas dapat mengembangkan strategi evaluasi dan perencanaan yang lebih baik untuk meningkatkan kelulusan mahasiswa, sekaligus memberikan dasar bagi pengembangan model prediksi yang lebih kompleks untuk institusi pendidikan lainnya.
Analysis of Public Sentiment Towards The TikTok Application Using The Naive Bayes Algorithm and Support Vector Machine Hidayah, Ika Arofatul Hidayah; Ririen Kusumawati; Zainal Abidin; M. Imamuddin
Journal of Computer Networks, Architecture and High Performance Computing Vol. 6 No. 2 (2024): Articles Research Volume 6 Issue 2, April 2024
Publisher : Information Technology and Science (ITScience)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47709/cnahpc.v6i2.3990

Abstract

In the current digital era, social media applications such as TikTok have become an important aspect of people's lives. TikTok allows users to create and share short videos, making it a global phenomenon with millions of active users. However, this application has also been the subject of various responses and opinions from the public. This research aims to classify public sentiment towards the TikTok application based on comments on Playstore using the Naïve Bayes algorithm and Support Vector Machine (SVM). This research method involves collecting comment data from Playstore using scraping techniques, resulting in 5,000 review data. Data pre-processing stages include case folding, tokenization, normalization, stopword removal, stemming, and data labeling using a lexicon. The data that has been processed is then weighted using Term Frequency - Inverse Document Frequency (TF-IDF) before being classified using the Naïve Bayes and SVM algorithms. Algorithm performance evaluation is carried out using the Confusion Matrix to measure accuracy, precision and recall. The research results show that the SVM algorithm has higher accuracy (84%) compared to Naïve Bayes (79%). SVM also shows better precision and recall values in classifying positive and negative sentiment from user reviews. From the results of the tests that have been carried out, the SVM algorithm is more effective than Naïve Bayes in sentiment analysis of the TikTok application. This research provides insight into how public sentiment can be measured and analyzed, and underscores the importance of choosing the right algorithm for data sentiment analysis on social media platforms.
Analisis Data Mining Untuk Deteksi Diabetes Mellitus Menggunakan Naïve Bayes Yusril Haza Mahendra; Ririen Kusumawati; Imamudin
Computer Science and Information Technology Vol 6 No 1 (2025): Jurnal Computer Science and Information Technology (CoSciTech)
Publisher : Universitas Muhammadiyah Riau

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37859/coscitech.v6i1.7954

Abstract

Diabetes mellitus is a chronic disease with a continuously increasing global prevalence. it is characterized by high blood glucose levels due to the body's inability to produce or effectively use insulin. the widespread impact on individuals and communities underscores the importance of early detection and proper management. In the digital era, data mining analysis has become a crucial tool in healthcare, enabling the exploration and analysis of health data on a large scale to identify patterns and trends that are difficult to detect manually. in the context of detecting diabetes mellitus, data mining holds great potential for predictive model development. one of the algorithms used is naïve bayes. This study analyzes naïve bayes classification for early symptoms of diabetes mellitus, with the aim of enhancing understanding of risk factors and developing early detection tools. The research findings indicate that naïve bayes has the highest accuracy of 78% with the application of missing value imputation mean. It is hoped that this research will enhance efforts in preventing and managing diabetes mellitus, as well as reducing the burden on individuals and communities as a whole.
Enhancing Student Collaboration in Academic Projects Through a Content-Based Filtering Recommender System Anwar, Aldian Faizzul; Kusumawati, Ririen; Yaqin, M. Ainul; Santoso, Irwan Budi; Zuhri, Abdurrozaq Ashshiddiqi
International Journal of Advances in Data and Information Systems Vol. 6 No. 2 (2025): August 2025 - International Journal of Advances in Data and Information Systems
Publisher : Indonesian Scientific Journal

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59395/ijadis.v6i2.1329

Abstract

The Informatics Engineering Study Program at UIN Maulana Malik Ibrahim Malang facilitates students in developing their interests and talents through 10 academic communities that serve as forums for knowledge exchange and innovation in IT project development. However, a challenge arises in assigning suitable students to appropriate projects, resulting in many projects being completed by a limited set of students. To address this, a recommender system for academic project members was developed using the Content-Based Filtering method. This system assists project initiators in selecting competent team members based on students’ prior experiences, considering the similarity between project requirements and student profiles. A dataset of 198 student-completed projects was used, with preprocessing, TF-IDF, and cosine similarity applied in the recommendation process. The system was implemented using the Flask framework with Python and HTML. Evaluation was conducted using the SUS method for usability (achieving a score of 79, categorized as excellent) and MAP for model performance across three scenarios. Scenario one (random community) scored 0.92, scenario two (same community) scored 0.79, and scenario three (comparison with actual members) scored 0.98. The results indicate that broader search scopes yield more accurate recommendations. This research contributes to the improvement of collaborative IT project in academic environments by enabling data-driven student member selection. The proposed system has the potential to be adopted by other academic institutions facing similar team formation challenges.
PENILAIAN KINERJA PEGAWAI DENGAN METODE TOPSIS DAN BACKPROPAGATION NEURAL NETWORK Yuliawan, Audi Bayu; Hariyadi, M. Amin; Kusumawati, Ririen; Crysdian, Cahyo; Nugroho, Fresy
JIPI (Jurnal Ilmiah Penelitian dan Pembelajaran Informatika) Vol 10, No 2 (2025)
Publisher : STKIP PGRI Tulungagung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29100/jipi.v10i2.7826

Abstract

Transformasi digital melalui penerapan Industri 4.0 dan e-Government telah mengubah paradigma administrasi publik, sehingga menuntut sistem evaluasi kinerja pegawai yang lebih adaptif dan objektif. Penelitian ini bertujuan untuk mengklasifikasikan kinerja pegawai ke dalam lima kategori, yaitu "sangat baik", "baik", "cukup", "buruk", dan "sangat buruk", dengan menggunakan pendekatan Neural Network Backpropagation. Metodologi yang digunakan mencakup beberapa tahapan utama, dimulai dari proses preprocessing data yang menge-lompokkan kriteria penilaian ke dalam empat aspek: kualifikasi, kom-petensi, kinerja, dan disiplin. Selanjutnya, dilakukan seleksi fitur menggunakan metode Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), dan hasilnya digunakan sebagai data pelatihan pada model Neural Network Backpropagation. Hasil pelati-han menunjukkan performa model yang cukup baik, dengan nilai loss dan Mean Squared Error (MSE) sebesar 0,000465, Mean Absolute Per-centage Error (MAPE) sebesar 19,59%, dan akurasi mencapai 80,41%. Sementara itu, hasil eksperimen dengan metode TOPSIS secara terpisah mencatat akurasi sebesar 81% dan nilai loss sebesar 0,377. Kombinasi metode TOPSIS dan Neural Network Backpropagation ter-bukti efektif dalam mengklasifikasikan kinerja pegawai secara konsis-ten. Temuan ini memberikan kontribusi terhadap pengembangan sis-tem evaluasi kinerja berbasis kecerdasan buatan yang lebih akurat dan adaptif terhadap tantangan administrasi publik modern.
KLASIFIKASI BERITA HOAKS BAHASA INDONESIA MENGGUNAKAN INDOBERT FINE-TUNING DENGAN PENDEKA-TAN FOCAL LOSS PADA DATA TIDAK SEIMBANG Kunaefi, Aang; Abidin, Zainal; Kusumawati, Ririen
JIPI (Jurnal Ilmiah Penelitian dan Pembelajaran Informatika) Vol 10, No 2 (2025)
Publisher : STKIP PGRI Tulungagung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29100/jipi.v10i2.7811

Abstract

Penyebaran berita hoaks di media online menjadi isu serius di tengah meningkatnya konsumsi informasi digital di kalangan masyarakat. Klasifikasi berita hoaks berbahasa Indonesia memiliki peran penting untuk menekan penyebaran informasi palsu. Salah satu tantangan utama dalam sistem klasifikasi ini adalah ketidakseimbangan distribusi data, di mana jumlah berita non-hoaks jauh lebih banyak dibanding-kan berita hoaks. Penelitian ini mengusulkan pendekatan klasifikasi berita hoaks berbahasa Indonesia melalui teknologi Natural Lan-guange Processing (NLP) menggunakan fine-tuning model IndoBERT, yang merupakan pre-trained language model berbasis arsitektur BERT (Bidirectional Encoder Representations from Transformers) dan dis-esuaikan untuk Bahasa Indonesia. Ketidakseimbangan data diatasi menggunakan metode Focal Loss. Pendekatan focal loss dirancang untuk lebih menekankan pembelajaran pada sampel kelas minoritas yang sulit diklasifikasikan. Penelitian ini menggunakan dataset dari platform Kaggle, Huggingfase dan Mendeley. Tataset mencakup berita Bahasa Indonesia dengan jumlah data berita hoaks jauh lebih kecil dari berita faktual. Hasil evaluasi menunjukkan bahwa kombinasi In-doBERT dan Focal Loss mampu meningkatkan performa model dengan akurasi sebesar 98.3% dibandingkan dengan pendekatan Cross-Entropy Loss yang mendapat akurasi 97% Penelitian ini menun-jukkan bahwa penggabungan model berbasis bahasa alami dengan strategi penanganan data tidak seimbang dapat memberikan hasil yang lebih akurat dalam mendeteksi berita hoaks.
Development of Academic Community Recommendation System Using Content-Based Filtering at UIN Malang Informatics Engineering Study Program Zuhri, Abdurrozzaaq Ashshiddiqi; Kusumawati, Ririen; Yaqin, Muhammad Ainul; Anwar, Aldian Faizzul; Pahlevi, Achmad Fahreza Alif
J-INTECH ( Journal of Information and Technology) Vol 13 No 01 (2025): J-Intech : Journal of Information and Technology
Publisher : LPPM STIKI MALANG

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32664/j-intech.v13i01.1916

Abstract

The mismatch between the number and quality of Information and Communication Technology (ICT) talents and industry needs in Indonesia creates significant challenges, especially for Informatics Engineering students who often experience difficulties in determining the appropriate professional field. This research aims to develop a content-based filtering-based academic community recommendation system to help students choose communities that are relevant to their interests, skills and experience. The system uses TF-IDF and cosine similarity methods to match student profiles with community descriptions. Data was collected from 48 students and 10 academic communities in the Informatics Engineering Study Program of UIN Malang, and processed through preprocessing stages before modeling. Evaluation results using the System Usability Scale (SUS) resulted in a score of 76, which is categorized in the “good” level, However, users indicated the need for improved guidance in navigating the system. This system is expected to be an innovative solution to increase student participation in appropriate academic communities, as well as support the development of their potential and readiness for the world of work
PREDICTION SYSTEM OF RICE CONSUMPTION NEEDS USING WEIGHTED MOVING AVERAGE METHOD Maulidifa, Renisa; Puspa Miladin Nuraida Safitri A.; Ririen Kusumawati
Jurnal Teknik Informatika (Jutif) Vol. 5 No. 4 (2024): JUTIF Volume 5, Number 4, August 2024
Publisher : Informatika, Universitas Jenderal Soedirman

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52436/1.jutif.2024.5.4.1905

Abstract

Most of Indonesia's population works in the agricultural sector. The main agricultural commodity is paddy which will be processed into rice. Despite being the fourth largest rice producer in the world, Indonesia continues to import rice. This is due to the rice deficit, declining rice field harvest areas, and the high consumption and demand for rice in the country. Malang Regency is one of the regions in Indonesia that faces challenges in fulfilling rice needs due to the increasing population and decreasing agricultural land due to land conversion. Therefore, this research aims to predict rice demand to ensure the availability of sufficient supply. This research implements the Weighted Moving Average (WMA) method to find the most optimal period and weight with the smallest MAPE value. The results show that WMA using a 3-month period and weights 0.1, 0.1, 0.8 is the best. From the test results, the rice demand obtained MAPE of 7.15% with the prediction results reaching 20,552.25 tons and the planting area obtained MAPE of 22.96% with the prediction results reaching 3842.70 ha for the next period. Further analysis was conducted to determine the efficiency of the available planting area whether it can sufficient the needs of rice. The results show that the expected rice production from the available planting area in Malang Regency can still sufficient the rice needs of the population. This research has also successfully implemented the method on a website-based system to facilitate data processing and prediction process with faster and more accurate results.
Utilizing the game design factor questionnaire to develop engaging games for adaptive learning in the serious educational game: the Ma'had Sari, Nur Fitriyah Ayu Tunjung; Kusumawati, Ririen; Karami, Ahmad Fahmi; A, Miftahul Hikmah Putri Samudera
OPSI Vol 17 No 1 (2024): ISSN 1693-2102
Publisher : Jurusan Teknik Industri, Fakultas Teknologi Industri UPN "Veteran" Yogyakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31315/opsi.v17i1.11322

Abstract

This study explores the unique mandatory residence, Ma'had Sunan Ampel Al Alyi (MSAA), at the State Islamic University (UIN) Maulana Malik Ibrahim Malang, aiming to develop Quranic reading competence in new students. The varying educational backgrounds of admitted students lead to differences in their Quranic reading capabilities, highlighting the need for adaptive learning. In response to this diversity, adaptive learning using artificial intelligence is employed, implemented through the serious education game "The Ma'had." Survey results from expert individuals using a Game Design Factor Questionnaire reveal the game's substantial potential. The results show high agreement (100%) on clear goals, engaging gameplay, and a sense of freedom, with 67% strongly agreeing on improved understanding. Challenges are motivating, and the game successfully sparks curiosity. "The Ma’had" Game proves effective, but further research is recommended to explore variations in player engagement and compare results with expert test subjects, employing alternative quantitative testing methods for a comprehensive analysis.
PEMETAAN SENTIMEN MASYARAKAT TERHADAP PILPRES 2024 DENGAN ALGORITMA SELF-ORGANIZING MAP Yuwono, Dwi Purbo; Santoso , Irwan Budi; Kusumawati, Ririen
Jurnal Review Pendidikan dan Pengajaran Vol. 7 No. 3 (2024): Volume 7 No 3 Tahun 2024
Publisher : LPPM Universitas Pahlawan Tuanku Tambusai

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31004/jrpp.v7i3.28830

Abstract

Pemilihan Umum (PEMILU) adalah salah satu cara untuk memilih presiden, kepala daerah, dan anggota parlemen yang berlangsung setiap lima tahun sekali. Dalam memasuki tahun- tahun politik saat ini akan banyak bertebaran informasi dan komentar dari masyarakat terhadap pelaksanaan pemilu, komentar atau pendapat yang disampaikan akan sangat beragam dimulai dari dukungan terhadap pelaksanaan pemilu, penggiringan opini publik, ujaran kebencian dan komentar-komentar lainnya. Kemajuan teknologi saat ini mengakibatkan penyampaian pendapat dapat dengan mudah dipublikasikan melalui media sosial, salah satunya adalah melalui media twitter, twitter menjadi salah satu media sosial yang paling sering digunakan masyarakat dalam mengemukakan pendapatnya karena dianggap bebas. Oleh karena itu, pada penelitian ini diusulakan pemataan sentimen atau opini masyarakat tentang Pilpres melalui X-Twitter, baik itu positif, negatif, atau netral dengan menggunakan Term Frequency-Inverse Document Frequency (TF-IDF) dan metode Self-Organizing Maps (SOM). Dari hasil penelitian didapatkan bahwa Algoritme TF-IDF dan Self-Organizing Maps (SOM) dengan sentimen cuitan pengguna Twitter dengan Hasil pengujian masing-masing model dengan menggunakan confusionmatrix didapatkan rata-rata accuracy sebesar 81%, precision 80,3%, recall 81%, dan f-measure 80%.