Claim Missing Document
Check
Articles

Found 36 Documents
Search

ANALISIS PENGGUNAAN METODE UNDERSAMPLING CLUSTER CENTROIDS UNTUK MODEL KLASIFIKASI SVM DAN EKSTRAKSI FITUR BOW PADA ULASAN APLIKASI SIREKAP 2024 Hidayati Ramadhani, Novia; Azmi Verdikha, Naufal; Azhima Yoga Siswa, Taghfirul
JATI (Jurnal Mahasiswa Teknik Informatika) Vol. 9 No. 3 (2025): JATI Vol. 9 No. 3
Publisher : Institut Teknologi Nasional Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36040/jati.v9i3.13536

Abstract

Sistem Informasi Rekapitulasi (SIREKAP) digunakan sebagai aplikasi perhitungan suara pemilu. Namun, pada awal bulan februari 2024 aplikasi ini mendapat banyak ulasan negatif di platform google play store, dengan rata-rata peringkat 2,7 dari 5. Hal ini menunjukkan adanya ketidakseimbangan pada peringkat ulasan pengguna. Ketidakseimbangan ini dapat mempengaruhi hasil analisis dan evaluasi aplikasi secara keseluruhan. Untuk mengatasi ketidakseimbangan data pada ulasan penelitian in menggunakan algoritma machine learning. Penelitian ini bertujuan untuk mengklasifikasikan teks ulasan aplikasi SIREKAP 2024 menggunakan algoritma SVM dengan ekstraksi fitur BoW serta metode undersampling Cluster Centroids (CC) untuk menangani ketidakseimbangan data. Dataset yang digunakan berjumlah 8235 ulasan. Proses validasi dilakukan menggunakan teknik k-fold cross-validation dengan k=10 untuk memastikan performa secara konsisten pada berbagai subset data. Hasil penelitian menunjukkan bahwa undersampling CC efektif dalam menyeimbangkan distribusi kelas, dengan Imbalanced Ratio (IR) awal sebesar 21,02 berhasil menjadi 1. Selain itu nilai entropi meningkat dari 1,079 menjadi 1,609 yang menunjukkan peningkatan keragaman ketidakpastian data. Namun, meskipun distribusi kelas menjadi seimbang, rata rata nilai IBA mengalami penurunan dari 0,274 pada Skenario I menjadi 0,271 pada Skenario II. Penurunan ini mengindikasikan bahwa teknik CC dapat menyebabkan hilangnya informasi penting pada kelas mayoritas selama proses undersampling.
Analisis Model Klasifikasi Sentimen Publik Terhadap Kebijakan Keberlanjutan IKN Menggunakan BERT Sebagai Feature Extractor dan K-Nearest Neighbor (KNN) Fiqri, Mohammad Hiqmal; Rudiman, Rudiman; Verdikha, Naufal Azmi
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.8168

Abstract

This study aims to evaluate the performance of sentiment classification models for public opinions regarding the relocation of Indonesia’s new capital (IKN) using a combination of IndoBERT as a feature extractor and K-Nearest Neighbor (KNN) as a classifier. The dataset consisted of 1,274 YouTube comments related to IKN, which were annotated by an expert in sociology and text analysis. The preprocessing stage involved cleaning numbers, URLs, emojis, and punctuation, as well as removing stopwords using the Sastrawi library. IndoBERT produced 768-dimensional vector representations, which were then classified using KNN with k=5 and Euclidean distance. Evaluation with 5-fold cross validation achieved an accuracy of 73.31%. However, the recall for the positive class was relatively low (0.49), indicating challenges in detecting positive comments due to class imbalance (831 negative, 294 positive, 149 neutral). These findings suggest that the IndoBERT+KNN model performs well on majority classes but struggles with minority classes. The contribution of this research is to provide a critical analysis of the limitations of IndoBERT-based models in Indonesian sentiment classification and to recommend future directions, including data balancing and fine-tuning approaches.
Analisis Sentimen Twitter Atas Isu Hak Angket Menggunakan Pembobotan TF-IDF dan Algoritma SVM Fahrezi, Irqi Anbi; Rudiman; Nauval Azmi Verdikha
Sci-tech Journal Vol. 3 No. 2 (2024): Sci-Tech Journal (STJ) In Press
Publisher : MES Bogor

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.56709/stj.v3i2.526

Abstract

Social media has become an important platform for voicing public opinion. One of the most popular and frequently used social media is Twitter. Twitter is a popular social media in Indonesia for discussions on political issues. The topic that is being discussed is the "inquiry right" because of the alleged fraud that occurred in the 2024 elections. The alleged fraud in the 2024 elections raised issues related to the rolling of the right of inquiry aimed at finding out the oddity or fraud. Therefore, a method is needed to classify the opinion whether it is classified as a positive or negative sentiment. This research uses 1113 data obtained from Twitter social media by applying crawling techniques. The data goes through several preprocessing stages then feature extraction using Term Frequency-Inverse Document Frequency, split data, and Support Vector Machine algorithms. The test results using these stages obtained an accuracy of 75%, indicating that the applied method is effective in classifying public sentiment related to the inquiry right issue..  
Klasifikasi Kecelakaan Lalu Lintas Menggunakan Kombinasi Forward Selection, ADASYN, dan Random Forest Aspianur; Taghfirul Azhima Yoga Siswa; Naufal Azmi Verdikha
Buffer Informatika Vol. 11 No. 2 (2025): Buffer Informatika
Publisher : Department of Informatics Engineering, Faculty of Computer Science, University of Kuningan, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Kecelakaan lalu lintas menjadi penyebab utama kematian bagi kelompok usia 15-29 tahun, dengan lebih dari 1,3 juta kematian setiap tahunnya. di Indonesia, data dari korps lalu lintas polri menunjukkan bahwa pada tahun 2023 terjadi lebih dari 100 ribu kasus kecelakaan, dengan korban jiwa mencapai 25 ribu orang. Penelitian ini bertujuan untuk mengembangkan model klasifikasi tingkat keparahan kecelakaan lalu lintas dengan mengintegrasikan metode Adaptive Synthetic Sampling (ADASYN) dan Forward Selection ke dalam algoritma Random Forest. Data yang digunakan merupakan data kecelakaan dari Polresta Samarinda periode 2020–2024 yang terdiri dari 35 fitur. Proses penelitian mencakup tahapan data preprocessing, feature selection dengan Forward Selection, dan pembagian data testing dan training dengan 10k-fold validation. Permodelan menggunakan algoritma Random Forest, dan evaluasi model menggunakan confusion matrix untuk mencari akurasi, precision, recall, dan f1-score. Hasil penelitian menunjukkan fitur yang paling signifikan terhadap klasifikasi kecelakaan lalu lintas adalah cuaca, jumlah luka ringan, jumlah luka berat, dan jumlah meninggal dunia. Penerapan Random Forest tanpa penanganan ketidakseimbangan dan tanpa seleksi fitur hanya menghasilkan akurasi 79,26%, precision 29,03%, recall 34,62%, dan f1-score 31,58%. Setelah diterapkan ADASYN, metrik evaluasi meningkat signifikan menjadi akurasi 84,26%, precision 81,82%, recall 84,62%, dan f1-score 83,20%. Peningkatan lebih besar tercapai setelah seleksi fitur Forward Selection, menghasilkan akurasi akhir 95,28%, precision 94,23%, recall 96,15%, dan f1-score 95,18%.
METODE PEMBOBOTAN TF-IDF UNTUK KLASIFIKASI TEKS QUICK COUNT PEMILIHAN WAKIL PRESIDEN INDONESIA 2024 PADA X TWITTER DENGAN METODE SVM Pranata, Ricky Albin; Rudiman, Rudiman; Azmi Verdikha, Naufal
Jurnal Teknologi Informasi: Jurnal Keilmuan dan Aplikasi Bidang Teknik Informatika Vol. 18 No. 2 (2024): Jurnal Teknologi Informasi : Jurnal Keilmuan dan Aplikasi Bidang Teknik Inform
Publisher : Universitas Palangka Raya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47111/jti.v18i2.14934

Abstract

The 2024 Indonesian Vice Presidential Election Quick Count sparked diverse public reactions on X Twitter. The sheer volume and variety of expressed opinions complicate accurate sentiment identification and classification. This study aims to develop a text classification model using Support Vector Machine (SVM) to identify sentiment in election Quick Count-related tweets. Data was acquired through tweet collection, followed by pre-processing, word weighting using TF-IDF, and data splitting for model training and testing. Results indicated that the developed SVM model achieved 77.30% accuracy in tweet sentiment classification. The model's implementation is expected to aid in more effective information filtering and assist stakeholders in understanding public opinion more accurately.
KLASIFIKASI SENTIMEN X-TWITTER PERIHAL PEMINDAHAN IBU KOTA INDONESIA MENGGUNAKAN EKSTRAKSI FITUR TF-IDF DAN METODE SUPPORT VECTOR MACHINE (SVM) Wahyudi, Tri; Rudiman, Rudiman; Verdikha, Naufal Azmi
Jurnal Teknologi Informasi: Jurnal Keilmuan dan Aplikasi Bidang Teknik Informatika Vol. 18 No. 2 (2024): Jurnal Teknologi Informasi : Jurnal Keilmuan dan Aplikasi Bidang Teknik Inform
Publisher : Universitas Palangka Raya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47111/jti.v18i2.15015

Abstract

The classification model has reached the realm of sentiment classification to analyze user sentiment in providing comments. this research aims to classify sentiment regarding the topic of moving the capital city of Indonesia using the Support Vector Machine (SVM) method with TF-IDF weighting. SVM has its own advantages, namely to overcome complex problems in SVM classification using the kernel function. the kernel functions to transform input data into a high dimensional feature space, allowing linear separation of data more easily. there are 3 sentiment categories in this study, namely Negative, Neutral and Positive sentiment. to determine these 3 categories, researchers used expert labelling services. the purpose of this study using the SVM method and TF-IDF feature extraction is to find out and analyze the accuracy results obtained in processing sentiment data regarding the transfer of the capital city of Indonesia. The accuracy results obtained are 64%, this shows that the SVM method with TF-IDF weighting is able to classify sentiment data with fairly good results.
PERFORMANCE OF TEXT SIMILARITY ALGORITHMS FOR ESSAY ANSWER SCORING IN ONLINE EXAMINATIONS Susanto, Muhammad Riza Radyaka; Husni Thamrin; Naufal Azmi Verdikha
Jurnal Teknik Informatika (Jutif) Vol. 4 No. 6 (2023): JUTIF Volume 4, Number 6, Desember 2023
Publisher : Informatika, Universitas Jenderal Soedirman

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52436/1.jutif.2023.4.6.1025

Abstract

The purpose of assessment is to determine learning success. Exams with question descriptions have several advantages, including ease of preparation and the ability to reveal student comprehension and originality. The problem with space is that it takes time to fix. Therefore, it is important to develop algorithms and software that automatically evaluate space. With the help of this algorithm and this software, you can solve some exam and assessment problems. This study aims to investigate similarity algorithms that approximate human patterns in evaluating ambiguous answers. This study examines his five similarity algorithms, including TF-IDF and LSA. The data was a collection of correct answers with a total of 371 texts. The similarity algorithm's performance was compared with human correction results. Evaluation was performed using Root Mean Square Error (RMSE). This study shows that his TF-IDF algorithm like Jaccard has the lowest his RMSE compared to human judgement. However, the LSA algorithm tended better to follow human rating patterns for descriptive tests..
PEMBUATAN APLIKASI ABSENSI DAN DOORPRIZE PENGUNJUNG BAPPEDA BERBASIS WEB PADA EVENT KALTIM EXPO 2023 Nurdiansyah, Rendy; Takhta Perlawanan Putra Sinawang; Reza Andriyanti; Naufal Azmi Verdikha
Jurnal Gembira: Pengabdian Kepada Masyarakat Vol 1 No 06 (2023): DESEMBER 2023
Publisher : Media Inovasi Pendidikan dan Publikasi

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Badan Perencanaan Pembangunan Daerah (BAPPEDA) adalah lembaga penting dalam perencanaan dan pembangunan daerah di Indonesia. Keterlibatan BAPPEDA dalam Kaltim Expo, sebagai bagian dari peringatan Hari Ulang Tahun Republik Indonesia ke-78, melibatkan presentasi kinerja serta hasil pencapaian, serta penyelenggaraan kuis dan undian doorprize untuk menarik perhatian pengunjung. Dalam upaya meningkatkan pengalaman pengunjung, BAPPEDA membutuhkan penggunaan teknologi terkini dengan sistem absensi yang efisien dan menarik, serta integrasi hadiah-hadiah menarik ke dalam sistem tersebut. Untuk memenuhi kebutuhan tersebut, dilakukan perancangan sistem absensi dan aplikasi doorprize berbasis web. Hal ini bertujuan untuk mendapatkan data absensi pengunjung secara efisien, memudahkan proses, dan menciptakan pengalaman yang lebih menarik bagi pengunjung KALTIM EXPO 2023. Pengembangan sistem absensi dan doorprize ini menggunakan metode Software Development Life Cycle (SDLC) dengan pendekatan Agile, fokus pada kerja tim kolaboratif yang responsif terhadap perubahan. Penggunaan bahasa pemrograman HTML, PHP, CSS, JS, dan Bootstrap 5 digunakan untuk tampilan, serta MySQL sebagai basis data.
Multilayer Perceptron and TF-IDF in the Classification of Hate Speech on Twitter in Indonesian Syahrandi, Akmal; Latipah, Asslia Johar; Verdikha, Naufal Azmi
JSE Journal of Science and Engineering Vol. 2 No. 1 (2023): Journal of Science and Engineering
Publisher : LPPI Universitas Muhammadiyah Kalimantan Timur (UMKT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30650/jse.v1i1.3773

Abstract

Twitter nowadays is one of the popular social media which currently has over 300millions accounts, twitter is the rich source to learn about people’s opion and sentimental analysis. However, this also brings new problems where the practice of hate speech. This research classifies of hate speech on social media. Evaluation using dataset from previous research Ibrohim&Budi (2019), then using classification method Multilayer Perceptron which combined with feature extraction to be able to detect negations and weighting uses Term Frequency – Inverse Document Frequency (TF-IDF). Results show that the F1 score gives an accuracy rate of up to 74.51%. This research has a reasonably good effectiveness from combining the TF-IDF and Multilayer Perceptron methods, considering the results obtained from the F1 Score evaluation value.
Indonesian Automated Essay Scoring with Bag of Word and Support Vector Regression Verdikha, Naufal Azmi; Dwiagam, Junianda Haris; Hasudungan, Rofilde
JSE Journal of Science and Engineering Vol. 2 No. 2 (2024): Journal of Science and Engineering
Publisher : LPPI Universitas Muhammadiyah Kalimantan Timur (UMKT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30650/jse.v1i2.3841

Abstract

Essay is one of the test questions to measure students' understanding of learning. Respondents can organize the answers to each question in their own language style, so it takes time to make corrections. It takes a system that can assess essay answers automatically quickly and accurately. Auto Essay Scoring (AES) is a tool that can assign grades or scores to answers in the form of essays automatically. In giving grades automatically, AES requires machine learning with training data that contains answer data that has been given a value by the assessor. In this study, AES was used to assess the Indonesian language midterm exams using the Bag of Word extraction feature and using Support Vector Regression. The Root Mean Square Error value obtained when evaluating AES is 1.99.