Garuda - Garba Rujukan Digital

Article Per Year (5 Year)

All

Tematik : Jurnal Teknologi Informasi Komunikasi

tematik
Website

Published by POLITEKNIK LP3I BANDUNG

ISSN : 23559055 EISSN : 24433640 DOI : 10.38204

Core Subject : Science,

Computer Science & IT

TEMATIK - Jurnal Teknologi Informasi Dan Komunikasi merupakan jurnal ilmiah sebagai bentuk pengabdian dalam hal pengembangan bidang Teknologi Informasi Dan Komunikasi serta bidang terkait lainnya. TEMATIK - Jurnal Teknologi Informasi Dan Komunikasi diterbitkan oleh LPPM dan Program Studi Manajemen Informatika di Politeknik LP3I Bandung. Redaksi mengundang para dosen, peneliti dan professional dari dunia industri dan kerja untuk menulis karya ilmiah dan pengalaman praktis di lapangan terkait implementasi Informatika dan Komputer.

Arjuna Subject : Ilmu Komputer (semua kategori) - Sistem Informasi

Articles 277 Documents

24 25 26 27 28

Deteksi Serangan SQL Injection Berulang pada Log Apache menggunakan Semantic Embedding dan Machine Learning Rosidin; Dian Ade Kurnia; Dadang Sudrajat
TEMATIK Vol. 13 No. 1 (2026): Tematik : Jurnal Teknologi Informasi Komunikasi (e-Journal) - Juni 2026
Publisher : LPPM POLITEKNIK LP3I BANDUNG

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.38204/tematik.v13i1.2909

SQL Injection (SQLi) attacks are a major threat to web application security, with attackers frequently attempting to modify query patterns repeatedly to evade detection. This research aims to develop a method for detecting repeated SQL Injection attacks in Apache logs using a semantic embedding approach. The method employed includes preprocessing log data using Natural Language Processing (NLP) techniques, generating vector representations through semantic embedding, and classifying using machine learning algorithms. The novelty of this research lies in the application of semantic embedding on Apache log queries to identify recurring and polymorphic SQL Injection attacks based on contextual similarity rather than exact syntactic matching. Experimental results show that the Support Vector Machine (SVM) model achieved the best performance with an accuracy of 94.5%, precision of 92.0%, recall of 91.0%, and F1-score of 91.5%, outperforming conventional rule-based detection methods.. The results show that the proposed approach is capable of recognizing semantic similarities between queries despite syntactic differences, thereby improving detection accuracy and reducing false positive rates compared to rule-based methods. Therefore, this method is effective for large-scale log analysis and is capable of detecting repeated and polymorphic SQL Injection attack pattern in web application environments.

Klasterisasi Produksi Unggas Indonesia Menggunakan K-Means untuk Mendukung Pengambilan Keputusan Berbasis Data Henny Yulianti; Mochamad Agung Wibowo; Budi Warsito
TEMATIK Vol. 13 No. 1 (2026): Tematik : Jurnal Teknologi Informasi Komunikasi (e-Journal) - Juni 2026
Publisher : LPPM POLITEKNIK LP3I BANDUNG

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.38204/tematik.v13i1.2930

Data utilization in decision-making within Indonesia’s poultry sector remains suboptimal despite the availability of heterogeneous production data across provinces. This study aims to cluster multi-commodity poultry production to identify regional distribution patterns and support data-driven decision-making. Unlike previous studies that mainly focused on cluster exploration and visualization without systematic relevance to policy needs, this research integrates K-Means Clustering with the CRISP-DM framework to produce a more systematic and applicable analysis. The dataset comprises poultry production data from 34 provinces in Indonesia during 2012–2023, including broiler chickens, layer chickens, ducks, quails, and native chickens as the main variables. The findings show that the optimal number of clusters, determined using the Silhouette Score, is four: very high, high, medium, and low production clusters, reflecting significant interregional disparities. The clustering pattern indicates spatial inequality, with high-production regions largely concentrated on Java Island, although North Sumatra and West Sumatra outside Java also demonstrate very high production levels, while most other provinces remain in the medium-to-low production categories. These findings provide an operational regional segmentation model and establish a measurable basis for prioritizing interventions, optimizing distribution systems, and formulating data-driven policies in Indonesia’s national poultry industry.

Analisis Pola Aktivitas Fisik Selama Ibadah Haji Menggunakan Data Smartwatch Longitudinal: Studi Kasus 784 Hari Fat'hah Noor Prawita; Dicky Rachmat Iskandar; Riska Ana Gulang
TEMATIK Vol. 13 No. 1 (2026): Tematik : Jurnal Teknologi Informasi Komunikasi (e-Journal) - Juni 2026
Publisher : LPPM POLITEKNIK LP3I BANDUNG

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.38204/tematik.v13i1.2934

Continuous monitoring using wearable smartwatches enables objective assessment of physical activity patterns in extreme religious activity contexts. This study analyzes 784 days of longitudinal data (February 2024 to April 2026) from an Amazfit T-Rex Pro smartwatch, focusing on the Hajj 1446 H period (17 May to 27 June 2025). The research uses a single-subject case study design involving a healthy 41-year-old male adult. Variables include daily step count, resting heart rate, and sleep duration, analyzed using descriptive statistics, Z-score based anomaly detection (Z > 2.0), and phase comparisons (pre-Hajj, during Hajj, post-Hajj). During Hajj, mean daily steps reached 14,871 steps per day, about twice the pre-Hajj baseline (7,480 steps per day). The core manasik phase (8 to 13 Dhul-Hijjah) showed a mean of 30,840 steps per day, peaking at 42,312 steps on Tasyrik I (11 Dhul-Hijjah), while Wukuf day (9 Dhul-Hijjah) recorded the lowest value of 5,491 steps (Z = -0.38). Resting heart rate remained stable (mean = 62.4 bpm), and sleep duration decreased during the core manasik phase (mean = 4.1 hours). A total of 38 anomaly days were identified, mostly during Hajj, indicating substantial phase dependent physical demands and the potential of wearable data for health monitoring in large-scale religious mass gatherings.

Implementasi Clean Architecture pada Chatbot Cuaca Hiper-Lokal Berbasis Hybrid NLP di Wilayah Bali Md Wira Putra Dananjaya
TEMATIK Vol. 13 No. 1 (2026): Tematik : Jurnal Teknologi Informasi Komunikasi (e-Journal) - Juni 2026
Publisher : LPPM POLITEKNIK LP3I BANDUNG

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.38204/tematik.v13i1.2938

Accurate meteorological information is crucial for the tourism and agriculture sectors in Bali Province. However, global weather Application Programming Interfaces (APIs) often have limitations in recognizing hyper-local areas and lack natural language interaction capabilities. This study aims to develop the Bali Weather Bot, an intelligent Telegram-based assistant using Clean Architecture and Hybrid Natural Language Processing (NLP). The methodology combines the Gemini 1.5 Flash Large Language Model (LLM) for entity extraction with a Context-Aware Dictionary Mapping fallback mechanism to standardize local abbreviations and Indonesian temporal metaphors. Evaluation was conducted using a dataset of 250 test query scenarios evaluated through a confusion matrix and latency benchmarking. System evaluation demonstrates that this hybrid architecture successfully extracts spatial and temporal parameters with an overall accuracy of 82.80%. Quantitative evaluation shows a Precision of 92.00%, Recall of 89.22%, and an F1-score of 90.58%, while maintaining a 0% system crash rate. In conclusion, the hybrid approach effectively mitigates AI hallucination and enhances hyper-local data retrieval reliability in conversational interfaces.

Analisis Sentimen Publik Terhadap Program Makan Bergizi Gratis Di Media Sosial Berbasis Transfer Learning Andriyas Ariya Firmansyah; Budi Sutomo
TEMATIK Vol. 13 No. 1 (2026): Tematik : Jurnal Teknologi Informasi Komunikasi (e-Journal) - Juni 2026
Publisher : LPPM POLITEKNIK LP3I BANDUNG

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.38204/tematik.v13i1.2960

Program Makan Bergizi Gratis menjadi salah satu kebijakan yang banyak dibicarakan di media sosial karena berkaitan langsung dengan isu gizi, distribusi bantuan, dan pelaksanaan program pemerintah. Penelitian ini bertujuan menganalisis sentimen masyarakat terhadap Program Makan Bergizi Gratis berdasarkan percakapan di Twitter sekaligus menguji kinerja model RoBERTa dalam klasifikasi sentimen. Data penelitian diperoleh melalui teknik scraping Twitter pada periode Januari 2025 sampai Maret 2026 dan menghasilkan 3.566 tweet yang disimpan dalam format CSV. Data selanjutnya diproses melalui tahap pre-processing yang meliputi cleaning, case folding, normalisasi kata, tokenisasi, dan stopword removal. Setelah itu, teks diterjemahkan ke dalam bahasa Inggris dan diberi label sentimen menggunakan metode VADER berdasarkan nilai compound score yang dikategorikan menjadi tiga kelas: yaitu positif, negatif, dan netral. Model RoBERTa dikembangkan dengan Hugging Face dan PyTorch, dengan pembagian dataset untuk 70% data pelatihan dan 30% untuk pengujian, serta menerapkan early stopping dengan patience 5 dan maksimum 30 epoch. Temuan penelitian mengungkap bahwa sentimen positif mendominasi dengan 70,27%, diikuti sentimen negatif 22,04%, dan netral 7,68%. Dari sisi performa model, RoBERTa menghasilkan accuracy 0,8709 serta macro F1-score 0,7970, menunjukkan efektivitas RoBERTa sekaligus menyoroti kendala pada kelas netral yang sulit diklasifikasikan. Pencapaian ini membuktikan bahwa RoBERTa cukup andal untuk analisis sentimen pada data Twitter, meskipun klasifikasi sentimen netral masih perlu ditingkatkan.

Analisis Sentimen Ulasan by.U dengan Pelabelan Rating dan Leksikon Menggunakan Multinomial Naïve Bayes Fatihanursari Dikananda; Bani Nurhakim; Dian Ade Kurnia; Ahmad Rifai; Mugi Praseptiawan
TEMATIK Vol. 13 No. 1 (2026): Tematik : Jurnal Teknologi Informasi Komunikasi (e-Journal) - Juni 2026
Publisher : LPPM POLITEKNIK LP3I BANDUNG

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.38204/tematik.v13i1.2996

Perkembangan layanan telekomunikasi digital mendorong bertambahnya jumlah ulasan pengguna yang digunakan sebagai bahan informasi guna mendukung pengambilan keputusan berbasis data. Penelitian ini bertujuan menganalisis sentimen ulasan aplikasi by.U menggunakan dua metode pelabelan data, yaitu rating-based labeling dan lexicon-based labeling, menggunakan algoritma Multinomial Naïve Bayes (MNB). Metode penelitian menerapkan framework Knowledge Discovery in Databases yang meliputi tahapan selection, preprocessing, transformation, data mining, dan evaluation. Dataset penelitian diperoleh dari Google Play sebanyak 8.000 ulasan berbahasa Indonesia. Tahap prapemrosesan mencakup cleaning, case folding, normalisasi, tokenisasi, stopword removal, serta stemming. Representasi fitur dilakukan menggunakan TF-IDF, sedangkan penyeimbangan data diterapkan melalui metode SMOTE. Evaluasi model dilakukan menggunakan metrik accuracy, precision, recall, dan F1-score dengan skema 10-fold cross validation. Hasil penelitian menunjukkan bahwa pendekatan lexicon-based labeling memberikan performa yang lebih baik dibandingkan rating-based labeling. Pendekatan rating-based menghasilkan accuracy sebesar 82,59%, precision 83,79%, recall 82,59%, dan F1-score 82,43%. Sementara itu, pendekatan lexicon-based memperoleh accuracy sebesar 88,96%, precision 89,69%, recall 88,96%, serta F1-score 88,91%. Temuan tersebut menunjukkan bahwa strategi pelabelan memiliki pengaruh terhadap performa klasifikasi sentimen. Pendekatan berbasis leksikon dinilai lebih efektif karena mampu memahami konteks linguistik dan ekspresi emosional pengguna secara lebih baik dibandingkan pendekatan berbasis rating.

Text Analysis of Workplace Accident Chronology for Hazard Identification Using BERT and Random Forest Algorithms Mohammad Zoqi Sarwani; Anang Aris Widodo
TEMATIK Vol. 13 No. 1 (2026): Tematik : Jurnal Teknologi Informasi Komunikasi (e-Journal) - Juni 2026
Publisher : LPPM POLITEKNIK LP3I BANDUNG

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.38204/tematik.v13i1.3004

Workplace accidents in industrial environments continue to cause significant human, economic, and operational losses, making proactive hazard identification a critical priority for occupational health and safety (OHS) management. This study develops a workplace hazard-identification pipeline that automatically analyzes accident chronology texts using BERT (Bidirectional Encoder Representations from Transformers) as a contextual feature extractor and Random Forest as an ensemble classifier. The publicly available IHMStefanini Industrial Safety and Health dataset was used as the source corpus; after data-quality screening, stratified train-validation-test splitting (80:10:10), and class-imbalance handling using SMOTE applied only to BERT embedding vectors in the training partition, the final modeling matrix covered 14 Critical Risk hazard categories. Text preprocessing included controlled normalization, tokenization with bert-base-uncased, padding and truncation to 128 tokens, and contextual embedding extraction into 768-dimensional feature vectors. Experimental results on the held-out test set show that the proposed BERT-Random Forest model achieved an accuracy of 94.7%, precision of 93.8%, recall of 94.2%, and F1-score of 94.0%, outperforming TF-IDF with SVM, Word2Vec with LSTM, BERT with SVM, and standalone BERT fine-tuning baselines. Statistical comparison using the McNemar-Bowker paired error test confirmed that the performance difference between the proposed model and the strongest baseline was significant (p < 0.01). The main contribution of this study is not the generic superiority of a hybrid BERT-Random Forest architecture, but its practical adaptation for multi-class workplace hazard identification from industrial accident narratives with lower computational cost than full transformer fine-tuning. The proposed method can support automatic incident triage, hazard monitoring, prioritization of safety investigations, and decision.

Page 28 of 28 | Total Record : 277

24 25 26 27 28

Filter by Year

2014 2026

Filter By Issues

Home Page

OAI Link

Editorial Team

Contact

Reviewer

Google Scholar

Contact Name
Hadiansyah

Contact Email
kanghadiansyah@plb.ac.id

Phone
+6285220199772

Journal Mail Official
tematik@plb.ac.id

Editorial Address
Program Studi Manajemen Informatika Politeknik LP3I Bandung Jl. Pahlawan No. 59 Bandung 40123 Telp. (022) 2506500, Fax. (022) 2512564 Email : tematik@plb.ac.id

Location
Kota bandung,

Jawa barat

INDONESIA

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Home Page

OAI Link

Editorial Team

Contact

Reviewer

Google Scholar

Contact Name Hadiansyah

Contact Email kanghadiansyah@plb.ac.id

Phone +6285220199772

Journal Mail Official tematik@plb.ac.id

Editorial Address Program Studi Manajemen Informatika Politeknik LP3I Bandung Jl. Pahlawan No. 59 Bandung 40123 Telp. (022) 2506500, Fax. (022) 2512564 Email : tematik@plb.ac.id

Location Kota bandung, Jawa barat INDONESIA

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Contact Name
Hadiansyah

Contact Email
kanghadiansyah@plb.ac.id

Phone
+6285220199772

Journal Mail Official
tematik@plb.ac.id

Editorial Address
Program Studi Manajemen Informatika Politeknik LP3I Bandung Jl. Pahlawan No. 59 Bandung 40123 Telp. (022) 2506500, Fax. (022) 2512564 Email : tematik@plb.ac.id

Location
Kota bandung,

Jawa barat

INDONESIA