Jurnal Online Informatika
Not a member yet
276 research outputs found
Sort by
Implementation of Generative Adversarial Network to Generate Fake Face Image
Full text linkIn recent years, many crimes use technology to generate someone\u27s face which has a bad effect on that person. Generative adversarial network is a method to generate fake images using discriminators and generators. Conventional GAN involved binary cross entropy loss for discriminator training to classify original image from dataset and fake image that generated from generator. However, use of binary cross entropy loss cannot provided gradient information to generator in creating a good fake image. When generator creates a fake image, discriminator only gives a little feedback (gradient information) to generator update its model. It causes generator take a long time to update the model. To solve this problem, there is an LSGAN that used a loss function (least squared loss). Discriminator can provide astrong gradient signal to generator update the model even though image was far from decision boundary. In making fake images, researchers used Least Squares GAN (LSGAN) with discriminator-1 loss value is 0.0061, discriminator-2 loss value is 0.0036, and generator loss value is 0.575. With the small loss value of the three important components, discriminator accuracy value in terms of classification reaches 95% for original image and 99% for fake image. In classified original image and fake image in this studyusing a supervised contrastive loss classification model with an accuracy value of 99.93%
User Experience Design and Prototypes of Mobile-based Learning Media for Children with Special Needs in the Dyslexia Category
Full text linkEducation is the right of all living things regardless of social status, gender, or physical condition. Persons with disabilities have the same rights and obligations as citizens. Based on the 1945 Constitution Article 31 Paragraph 1 and Law Number 20 of 2003 concerning the National Education System, it can be concluded that the state provides full guarantees for children with special needs to obtain quality education services. Children with special needs are divided into several categories, in this study the research team will focus on solving learning problems for children with disabilities in the dyslexia category. Dyslexia also known as reading disorder, is a disorder characterized by reading below the expected level for one\u27s age. This study aims to find learning solutions by developing user experience designs and prototypes of mobile-based learning media for children with special needs in the dyslexia category. This research applies design thinking methodology to understand users, challenge assumptions, redefine problems, and create innovative solutions to prototype and test
Identifikasi Kesamaan Pertanyaan pada Soal Bahasa Indonesia Menggunakan Metode Recurrent Neural Network (RNN)
No full textIn a question-and-answer forum, the identification of question similarity is used to determine how similar two questions are. This procedure makes sure that user-submitted questions are compared to the questions in a database for matches to improve system performance on the online Q&A platform. Currently, question similarity is mostly done in foreign languages. The purpose of this research is to identify question similarities and evaluate the effectiveness of the methods used in Indonesian language questions. The data used is a public dataset with labeled pairs of questions as 0 and 1 where label 0 for different pairs of questions and label 1 for the same pairs of questions. The method used is a Recurrent Neural Network (RNN) with the Manhattan Distance approach to calculate the similarity distance between two questions. The question pairs are taken as two inputs with a reference label to identify the similarity distance between the two question inputs. We evaluated the model using three different optimizers namely RMSprop, Adam, and Adagrad. The best results were obtained using the Adam optimizer with 80:20 ratio split-data and overall accuracy is 76%, precision is 74%, recall is 98.8%, and F1-score is 85.1%.Identifikasi kesamaan pertanyaan merupakan bagian penting dalam Question Answering System. Identifikasi kesamaan pertanyaan dilakukan dengan tujuan untuk membuat sebuah sistem menjadi lebih efisien dalam memberikan jawaban secara cepat dan akurat. Fokus yang dikerjakan pada penelitian ini adalah mengidentifikasi kesamaan pertanyaan pada soal bahasa Indonesia serta mengevaluasi efektivitas penggunaan metode pada bahasa Indonesia. Data yang digunakan adalah dataset dengan pasangan pertanyaan berlabel 0 dan 1 dimana label 0 untuk pasangan pertanyaan yang berbeda dan label 1 untuk pasangan pertanyaan yang sama. Kesamaan pertanyaan tersebut diidetifikasi dengan menggunakan model Recurrent Neural Network (RNN) dengan pendeketakan Manhattan Distance. Pasangan pertanyaan dijadikan sebagai dua inputan dengan acuan label untuk mengidentifikasi jarak kesamaan antara kedua inputan tersebut menggunakan pendekatan Manhattan distance. Model dievaluasi dan menghasilkan skor akurasi sebesar 76%, presisi 74%, recall 98,8% dan f1-score 85,1%. Hasil tersebut diperoleh melalui penambahan stopword_removal pada data. Analisis kami terhadap hasil yang didapatkan adalah dengan penambahan fungsi stopword_removal pada proses preprocessing dapat meningkatkan hasil identifikasi kesamaan pertanyaan pada soal bahasa Indonesia
Automate IGP and EGP Routing Protocol Configuration using a Network Automation Library
Full text linkData communication is sending data from client to client through a computer network. The increasing use of data communication makes computer networks more complex. Complex computer networks make it difficult for network administrators to configure them, especially routing protocol configuration. Network administrators are in charge of configuring routing protocols and managing networks. In addition, the more devices on the network, the greater the chance of human error from the administrator. Therefore, network automation is one solution that helps network administrators overcome this. This study focuses on analyzing the performance of network automation using the Paramiko and Telnetlib libraries. The routing protocol used by OSPF for IGP and BGP for EGP. The scenario in this study involves configuring IP addresses and configuring OSPF and BGP routing. Based on the test results, the Telnetlib library is better than the Paramiko library in terms of script delivery time, convergence time, and delay by 19.237% when applied to the IGP and EGP routing protocols
Retweet Prediction Using Multi-Layer Perceptron Optimized by The Swarm Intelligence Algorithm
Full text linkRetweets are a way to spread information on Twitter. A tweet is affected by several features which determine whether a tweet will be retweeted or not. In this research, we discuss the features that influence the spread of a tweet. These features are user-based, time-based and content-based. User-based features are related to the user who tweeted, time-based features are related to when the tweet was uploaded, while content-based features are features related to the content of the tweet. The classifier used to predict whether a tweet will be retweeted is Multi Layer Perceptron (MLP) and MLP which is optimized by the swarm intelligence algorithm. In this research, data from Indonesian Twitter users with the hashtag FIFA U-20 was used. The results of this research show that the most influential feature in determining whether a tweet will be retweeted or not is the content-based feature. Furthermore, it was found that the MLP optimized with the swarm intelligence algorithm had better performance compared to the MLP
Implementation of Dynamic Topic Modeling to Discover Topic Evolution on Customer Reviews
Full text linkAnnotation and analysis of online customer reviews were identified as significant problems in various domains, including business intelligence, marketing, and e-governance. In the last decade, various approaches based on topic modeling have been developed to solve this problem. The known solutions, however, often only work well on content with static topics. As a result, it is challenging to analyze customer reviews that include dynamic and constantly expanding collections of short and noisy texts. A method was proposed to handle such dynamic content. The proposed system applied a dynamic topic model using BERTopic to monitor topics and word evolution over time. It would help decide when the topic model needs to be retrained to capture emerging topics. Several experiments were conducted to test the practicality and effectiveness of the proposed framework. It demonstrated how a dynamic topic model could handle the emergence of new and over-time-correlated topics in customer review data. As a result, improved performance was achieved compared to the baseline static topic model, with 25% of new segmented texts discovered using the dynamic topic model. Experimental results have, therefore, convincingly demonstrated that the proposed framework can be used in practice to develop automatic review annotation tools
Multi-Step Vector Output Prediction of Time Series Using EMA LSTM
Full text linkThis research paper proposes a novel method, Exponential Moving Average Long Short-Term Memory (EMA LSTM), for multi-step vector output prediction of time series data using deep learning. The method combines the LSTM with the exponential moving average (EMA) technique to reduce noise in the data and improve the accuracy of prediction. The research compares the performance of EMA LSTM to other commonly used deep learning models, including LSTM, GRU, RNN, and CNN, and evaluates the results using statistical tests. The dataset used in this study contains daily stock market prices for several years, with inputs of 60, 90, and 120 previous days, and predictions for the next 20 and 30 days. The results show that the EMA LSTM method outperforms other models in terms of accuracy, with lower RMSE and MAPE values. This study has important implications for real-world applications, such as stock market forecasting and climate prediction, and highlights the importance of careful preprocessing of the data to improve the performance of deep learning models
The Impact of Data Augmentation Techniques on the Recognition of Script Images in Deep Learning Models
Full text linkDeep learning technology is widely used for recognizing character images, including various regional characters and diverse ancient scripts. Deep learning models require large data sets to recognize images accurately. However, creating a dataset has limitations in terms of quantity, including the Komering script dataset used in this study. Data augmentation techniques can be applied to expand the dataset by modifying existing images to increase data diversity. This study aims to investigate the impact of augmentation techniques on the performance of deep learning models in the case of Komering script recognition. The dataset consists of 500 images for five classes of Komering script characters. Three augmentation techniques, namely random rotation, height shift, and width shift, were applied to the five characters, which were then used to test the model trained to recognize characters in the Komering dataset. This research contributes to providing insights into the effect of augmentation techniques on robust confidence prediction of deep learning models for recognizing newly augmented data. The results demonstrate that the deep learning model can recognize modified data using augmentation techniques with an average accuracy of 80.05%.Aksara komering merupakan aksara daerah yang dimiliki oleh suku Komering di Sumatera Selatan. Aksara ini memilki beragam karakter yang hampir sama bentuknya. Untuk mempermudah mengenali gambar aksara tersebut maka bisa digunakan teknologi deep learning. Model deep learning sendiri memerlukan dataset dalam jumlah banyak untuk bisa mengenali gambar secara tepat. Pembuatan dataset sendiri biasanya memiliki keterbatasan jumlah, demikian juga untuk dataset aksara Komering yang digunakan dalam penelitian ini. Untuk memperbanyak dataset bisa dilakukan dengan teknik augmentasi untuk menjadikan data lebih beragam. Teknik augmentasi merupakan salah satu cara yang bisa digunakan untuk menambah jumlah data dengan berbagai teknik memodifikasi gambar yang sudah ada. Penelitian ini bertujuan untuk mengetahui pengaruh teknik augmentasi terhadap kinerja pada model Deep Learning pengenalan Aksara Komering. Ada tiga teknik augumentasi yang dilakukan yaitu random rotation, height shift, dan width shift. Teknik augmentasi dilakukan pada 5 karakter yang kemudian digunakan untuk data uji model yang sudah dilatih mengenali karakter pada dataset Komering. Hasilnya memperlihatkan data augmentasi yang sudah dimodifikasi dengan width shift menghasilkan kinerja yang lebih baik
XGBoost and Convolutional Neural Network Classification Models on Pronunciation of Hijaiyah Letters According to Sanad
Full text linkAccording to Sanad, the pronunciation of Hijaiyah letters can serve as a benchmark for correct or valid reading based on the makhraj and properties of the letters. However, the limited number of Qur\u27anic Sanad teachers remains one of the obstacles to learning the Qur\u27an. This study aims to identify the most practical combination of classification models in constructing a voice recognition system that facilitates learning without requiring direct interaction with a teacher. The methods employed include the XGBoost algorithm and CNN. As a result, out of the 12 letter trait labels, the CNN model was utilized for 10 of them, specifically for traits S1, S2, S4, S5, T1, T2, T3, T4, T5, and T6, on trait labels S3 and T7 applying the XGBoost model. Furthermore, the inclusion of additional data yielded performance results for each property, with an average accuracy of 78.14% for property S (letters with opposing properties), 70.69% for property T (letters without opposing properties), and an overall average of 73.79% per letter.Huruf Hijaiyah adalah huruf yang terdapat dalam susunan Al-Qur\u27an. Karakter huruf hijaiyah adalah kenampakan karakter yang keluar dari makhrajnya, sedangkan huruf makharijul adalah tempat keluarnya huruf saat melafalkan huruf hijaiyah. Huruf Hijaiyah yang ada dalam Sanad dapat dijadikan sebagai patokan bacaan yang benar atau sahih karena sudah memenuhi ciri dan makna huruf tersebut. Terbatasnya jumlah pengajar Al-Qur\u27an masih menjadi salah satu kendala dalam mempelajari Al-Qur\u27an dengan baik. Hal ini ditunjukkan dengan sedikitnya pengajaran Al-Qur\u27an pada Sanad yang membuka pembelajaran tahsin, padahal pembelajaran tahsin pada Sanad merupakan salah satu pelajaran yang memiliki standar dalam pengucapan kaidah huruf sesuai dengan sifat hurufnya. Sistem pengenalan suara dapat mengenali suara sehingga dengan teknologi ini diharapkan dapat mendukung pembelajaran tanpa harus bertemu dengan guru. Pada penelitian ini dibangun model klasifikasi huruf hijaiyah berdasarkan karakteristik huruf menggunakan algoritma pembelajaran dangkal XGBoost dan algoritma pembelajaran mendalam CNN. CNN cenderung menghasilkan kinerja yang lebih baik daripada Extreme Gradient Boosting (XGBoost). Model algoritme XGBoost memiliki akurasi superior pada properti S2 dan T7. Namun, memorinya rendah. Penambahan data memberikan keseimbangan terhadap hasil kinerja sehingga nilai akurasi, presisi, memori, dan skor F-1 memiliki tingkat yang cukup. Akurasi untuk sifat S rata-rata 78,14%, properti T 70,69%, dan rata-rata per huruf 73,79%
Improving Indonesian Named Entity Recognition for Domain Zakat Using Conditional Random Fields
Full text linkIn Indonesia, where the majority of the population is Muslim, one of the obligations of a Muslim is zakat. To reduce illiteracy about zakat among Muslims, they need to have access to basic information about it. In order to facilitate the acquisition of this information, this study utilized named entity recognition (NER) and defined 12 named entity classes for the zakat domain, including the pillars of Islam, various types of zakat, and zakat management institutions. The Conditional Random Fields method was used for testing Indonesian-NER in three scenarios. In the specific context of the Zakat domain, NER can extract information about organizations, individuals, and locations involved in collecting and distributing Zakat funds. This information can improve the Zakat system’s efficiency and transparency and support research and analysis on Zakat-related topics. The average performance evaluation of the Indonesian-NER model showed a precision of 0.902, recall of 0.834, and an F1-score of 0.867