Jurnal Informatika: Jurnal Pengembangan IT
Not a member yet
437 research outputs found
Sort by
Perbandingan Cosine Similarity dan Weighted Jaccard Similarity dalam Pengembangan Mesin Pencari Perpustakaan Digital
No full textThis study addressed the problem of low relevance in search results within the digital library system of the Department of Informatics and Computer Engineering (JTIK), Universitas Negeri Makassar. The purpose of this research was to improve the accuracy and relevance of search outcomes, enabling users, particularly students, to access academic materials and research references more efficiently. A search engine system was developed using a term-weighting method based on term frequency and document distribution. The system incorporated similarity measurement techniques to evaluate the degree of match between user queries and document content. An experimental approach was applied, which involved observation, data collection, text preprocessing, implementation of term weighting, and the comparison of cosine similarity and Weighted Jaccard similarity for ranking search results. The The evaluation was conducted using the Precision@K metric and a paired t-test to measure the significance of performance differences between methods. The test results showed that Weighted Jaccard obtained an average Precision@K value of 0.933, slightly higher than Cosine Similarity with an average of 0.9. However, Cosine Similarity produced a higher average similarity value. In addition, system testing was conducted in two stages, namely assessing user satisfaction with search results and assessing system performance. These findings confirmed that the combination of term-weighting and cosine similarity effectively enhanced the relevance and performance of digital library search systems
Pencarian Rute Terpendek untuk Pemetaan UMKM di Kecamatan Negeri Katon Menggunakan Algoritma A-Star
No full textIn the regional economy, micro, small and medium enterprises (MSMEs) have a very significant role in driving economic growth and opening up employment opportunities for the community. However, the lack of information regarding business locations and optimal routes for MSMEs is still an obstacle in improving accessibility. The lack of information regarding the fastest route to the location of MSMEs leads to limited accessibility, especially for people who are unfamiliar with the area. Therefore, a more optimal strategy is needed to determine the best way to improve the distribution efficiency and mobility of MSMEs. The purpose of this research is to use the A-Star algorithm for mapping MSMEs to find the fastest route to the location of MSMEs. This research explicitly combines MSME spatial data with the implementation of the A-Star algorithm for route optimization. The results show that the A-Star algorithm is able to effectively speed up the route search process by taking into account the appropriate heuristic value. With the implementation of this algorithm, accessibility to MSMEs locations is significantly improved, allowing customers and businesses to easily find MSME locations with greater cost and time efficiency. Implementation of the A-Star algorithm in increase in MSME accessibility through optimal route efficiency
Efektivitas Flashcard Digital Sebagai Media Pedagogis dalam Pembelajaran Kosakata Bahasa Inggris: Systematic Literature Review
No full textVocabulary mastery is a fundamental aspect in English language learning. However, vocabulary learning is often boring and less interactive, so students have difficulty remembering and understanding new words. Digital learning applications such as flashcards offer an innovative solution to overcome this obstacle by having interactive features that can improve students' memory and learning interest. This study aims to evaluate the effectiveness of using digital flashcard applications as a pedagogical medium for teaching English vocabulary. Although some findings state that flashcards are effective in improving vocabulary in learning activities, a systematic review of the existing literature is needed to assess the extent to which digital flashcard applications have an impact on vocabulary learning both in terms of improving learning outcomes, learning motivation, and other aspects related to pedagogy. This study uses the Systematic Literature Review (SLR) method by reviewing several journals discussing the research topic. The results of the study indicate that digital flashcards as a learning media strategy can be said to be effective in improving students' vocabulary mastery ranging from elementary school, MTs, vocational high schools, to undergraduate students, which in this case can enrich vocabulary; easy, easy to recognize images, words, and meanings; Students are more active, focused, and enthusiastic when learning English; improve receptive and productive aspects in vocabulary knowledge; remembering material easily; and increasing students' learning motivation
Optimasi AdaBoost dan XGBoost untuk Klasifikasi Obesitas Menggunakan SMOTE
No full textObesity is a condition in which a person's weight exceeds the normal limit due to excessive accumulation of fat tissue. Thus, obesity is considered a global public health challenge. This is evidenced by the latest data from the World Health Organization (WHO) in 2022, namely that 2.5 billion adults aged 18 years and over are overweight and 890 million of them are obese. Therefore, it is very important to accurately identify these risk factors in order to implement effective interventions in the prevention and management of obesity. However, in previous studies there has been no application of SMOTE with the AdaBoost and XGBoost algorithms, so this study aims to compare the performance of the AdaBoost and XGBoost algorithms with SMOTE. The stages of this research begin with problem identification, data collection, preprocessing and model evaluation and model comparison. This study also applies the SMOTE technique to balance unbalanced data. Based on the results of the research that has been carried out, it shows that the accuracy and recall values of the XGBoost algorithm with SMOTE are 0.945 and precision 0.947. Meanwhile, the accuracy and recall values on AdaBoost with SMOTE are 0.388. Then, the precision is 0.371. Thus, it is expected that the results of the XGBoost model with SMOTE can be a source for other research and can help in efforts to prevent and manage obesity
Analisis Berbasis Convolutional Neural Network untuk Pendeteksian Kanker Prostat dengan Citra Magnetic Resonance Imaging (MRI)
No full textKanker prostat adalah tumor ganas yang berada dari kelenjar prostat, yang merupakan bagian penting dari sistem reproduksi pria. Adanya peningkatan prevalensi kanker prostat maka diperlukan deteksi dini yang akurat. Penelitian ini memfokuskan pada pemanfaatan deep learning, khususnya metode Convolutional Neural Network (CNN) untuk mendiagnosis kanker prostat melalui citra MRI. Diperlukan penelitian untuk mengkaji tiga arsitektur CNN: U-Net, nnU-Net, dan nnDetection agar didapatkan arsitektur yang terbaik. Data penelitian ini menggunakan data sekunder sejumlah 1294 citra MRI dari The PI-CAI Challenge “Artificial Intelligence Radiologists Prostate Cancer Detection in MRI” tahun 2022. Data tersebut menjalani proses pre-processing, termasuk normalisasi intensitas piksel, augmentasi data seperti rotasi dan scaling, serta pemotongan gambar untuk menghilangkan area yang tidak relevan. Proses selanjutnya data tersebut akan masuk ke tahap pelatihan model dengan menggunakan ketiga arsitektur. Hasil dari pelatihan tersebut akan dievaluasi kinerja modelnya dengan menggunakan metrik Area Under the Receiver Operating Characteristic Curve (AUROC) dan Average Precision (AP). Hasil evaluasi menunjukkan bahwa arsitektur U-Net mencapai AUROC 89,94% dan AP 51,22%, arsitektur nnU-Net mencapai AUROC 97,75% dan AP 86,67%. dan arsitektur nnDetection mencapai AUROC 83,66% serta AP 49,91%. Dari perbandingan hasil ketiga arsitektur maka didapatkan hasil terbaik adalah arsitektur nnU-Net dengan capaian AUROC 97,75% dan AP 86,67%. Penelitian ini menunjukkan potensi penggunaan CNN dalam diagnosis kanker prostat melalui citra MRI. Temuan penelitian menegaskan pentingnya pemilihan arsitektur yang tepat dalam aplikasi deep learning untuk citra medis
Perbandingan IndoBERT dan IndoRoBERTa Untuk Analisis Sentimen Pada Film Dokumenter Dirty Vote
No full textSentiment analysis is a technique in Natural Language Processing (NLP) used to identify and categorize opinions or emotions in text. This study compares the performance of two Transformer-based models, IndoBERT and IndoRoBERTa, in analyzing sentiment toward the documentary film Dirty Vote. The research process includes data collection, text preprocessing, lexicon-based sentiment labeling, and model evaluation using K-Fold Cross-Validation. The results show that IndoBERT achieved an average accuracy of 99%, higher than IndoRoBERTa, which achieved 94%. IndoBERT also demonstrated better alignment with lexicon-based labeling in classifying positive, negative, and neutral sentiments. In terms of architecture, IndoBERT employs static masking, while IndoRoBERTa applies dynamic masking, leading to differences in the models' sensitivity to textual meaning. IndoBERT tends to provide more definitive classifications for opinions or strong criticisms, whereas IndoRoBERTa more frequently categorizes ambiguous comments as neutral sentiment. The conclusion of this study indicates that IndoBERT outperforms IndoRoBERTa in sentiment analysis of the documentary film Dirty Vote, both in terms of accuracy and consistency with lexicon-based labeling. These findings provide insights into the effectiveness of Transformer-based models for sentiment analysis in the Indonesian language and can serve as a reference for further NLP model development
SCANOCULAR: Application for Early Detection of Eye Diseases Using AI and Blockchain Technology
No full textEye diseases such as cataracts, glaucoma, and diabetic retinopathy affect approximately 2.2 billion people globally, with 1 billion cases being preventable. In Indonesia, cataracts remain the leading cause of blindness. This research presents SCANOCULAR, a mobile application that integrates artificial intelligence (AI) and blockchain technology for early detection of eye diseases. The system utilizes a modified EfficientNetB4 Convolutional Neural Network (CNN) for analyzing eye images, achieving 95.50% accuracy, 95.92% precision, and 94.95% recall in cataract detection with an AUC of 0.9932. The blockchain implementation using Polygon Amoy platform ensures secure data transmission and storage while maintaining efficient transaction processing. Testing results demonstrate the system's capability in identifying various eye conditions while maintaining data integrity through blockchain verification. SCANOCULAR contributes to informatics by implementing a hybrid AI-blockchain architecture optimized for medical imaging applications, with a lightweight CNN model design that reduces computational requirements while maintaining diagnostic accuracy. This integration of technologies provides a potential solution for improving accessibility to eye disease screening and early intervention in Indonesia
Pengembangan Media Pembelajaran Pengenalan Perangkat Keras Jaringang Komputer Berbasis Augmented Reality
No full textIn Indonesia, many educators still apply conventional methods, namely by reading questions to students which are then answered using paper media, which are often considered less effective in attracting students' interest in learning. One of them is in the material on the introduction of computer network hardware. The purpose of this study is to create Augmented reality (AR)-based learning media for the introduction of computer network hardware in order to create a fun and non-boring learning process for students. The process of creating learning media is carried out by applying the Multimedia Development Life Cycle (MDLC) method. This study resulted in an application of computer network hardware learning media based on augmented reality. This application is equipped with learning materials, learning objectives, images and 3D objects. Based on the validation results, the application obtained a feasibility level of 86.12% from media experts, 84.45% from material experts, and 79.37% from user tests (students). These findings indicate that the application of AR not only increases visual appeal, but also supports conceptual understanding through immersive visual representations. The main contribution of this study lies in the systematic application of the MDLC framework in the development of AR learning media, as well as the integration of 3D objects designed according to pedagogical needs in the context of ICT education
Optimalisasi Portofolio Saham Syariah Berbasis Prediksi Menggunakan Long Short-Term Memory (LSTM)
No full textSaham merupakan salah satu jenis investasi aset finansial yang berpotensi untuk memberikan tingkat imbal balik yang tinggi sehingga menjadi salah satu instrument investasi yang popular. Salah satu jenis saham yang popular di Indonesia adalah saham syariah yang didukung kuat dengan ajaran agama islam (shariah compliant). Saham syariah mempunyai kinerja yang baik jika dibandingkan dengan saham konvensional ketika terjadi krisis keuangan ditandai dengan risiko indeks yang lebih kecil. Investor saham selalu menginginkan hasil timbal balik yang maksimal dengan risiko seminimal mungkin. Keinginan tersebut dapat tercapai dengan menyeleksi saham dengan return terbesar lalu melakukan optimalisasi pada potofolio saham. Salah satu metode seleksi saham yang dapat dilakukan adalah dengan memprediksi harga saham dengan menggunakan model LSTM pada indeks JII. Saham dengan return terbesar sesuai dengan hasil prediksi akan dimasukkan ke dalam satu portofolio yang akan dioptimalisasi dengan metode Mean-Variance (MV) dan Equal Weight (EW) yang akan diambil metode terbaik. Sebagai pembanding, portofolio dengan saham yang dipilih secara acak akan dibentuk dan dibandingkan hasilnya. Hasil penelitian menunjukkan portofolio yang dibentuk dengan menggunakan prediksi model LSTM dan metode optimalisasi MV memiliki keseimbangan dalam nilai mean return bulanan, standar deviasi bulanan, sharpe ratio bulanan, serta simulasi investasi sepanjang tahun 2023
Klasifikasi Pengucapan Huruf Hijaiyah Berbasis Android Menggunakan CNN dengan Fitur Mel-Spectrogram
No full textMastery of Hijaiyah letters is a fundamental basis in learning the Qur'an, but data from the IIQ Community Service Institute 2021/2022 shows that 72.25% of the 3,111 Muslims tested have not been able to read the Qur'an properly. This research aims to develop an Android-based Hijaiyah letter pronunciation classification system using Convolutional Neural Network (CNN) with mel-spectrogram features. The research methodology includes collecting 8,904 voice samples from 53 participants at Pondok Tahfidz Yanbu'ul Qur'an Menawan, preprocessing data using MFCC techniques, developing CNN models, and implementing the system in the form of mobile applications with MVVM architecture. The test results showed promising performance with some classes achieving 100% accuracy and an average overall accuracy of 83.80%, although there were challenges in some classes such as “alif_dommah” and “ghaiin_dommah” which had an accuracy below 40%. The developed system successfully provides an interactive learning platform through the integration of mobile applications with the Flask API, but still requires further development, especially in expanding the dataset to overcome overfitting problems and improve the generalization ability of the model