Articles
<![CDATA[Studi Literatur: Optimasi Algoritma Machine Learning Untuk Prediksi Penerimaan Mahasiswa Pascasarjana ]]>
<![CDATA[Zuhri, Burhanudin]]>;
<![CDATA[Harani, Nisa Hanum]]>
<![CDATA[ Jurnal Informatika dan Teknologi Komputer (J-ICOM) Vol 5 No 1 (2024): Jurnal Informatika dan Teknologi Komputer ( JICOM) ]]>
Publisher : <![CDATA[Universitas Samudra ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.55377/j-icom.v5i1.8074
<![CDATA[Machine learning algorithms are mathematical procedures used to find complex and hidden patterns in data with a high degree of accuracy and have brought major advances in various fields for fast and precise decision making. One of these fields is the field of education, which is to predict the admissions process for postgraduate students. The purpose of admitting postgraduate students is to select prospective students who are qualified and meet the academic requirements set by the institution concerned based on GRE (Graduate Record Examination) scores, TOEFL (Test of English as a Foreign Language) scores, university rankings, letters of recommendation, GPA bachelor degree, and research experience. Success in postgraduate admissions can open opportunities to earn advanced degrees and acquire more in-depth knowledge and skills in areas of interest. In this study, an analysis was carried out on various machine learning algorithm optimizations used to optimize topics or trends in previous studies. In this case, the researcher compares performance and selects the best algorithm optimization to be applied to the topic of graduate student admissions. The results of this review show that the hybrid algorithm has the best performance in optimizing predictions for most of the data in previous studies. The results of this study indicate that the CNN-LSTM (Convolutional Neural Network - Long Short-Term Memory) hybrid model is expected to be an appropriate alternative in optimizing predictions of postgraduate student admissions. Therefore, further research is needed to develop this algorithm and expand its application to the topic of graduate student admissions.]]>
<![CDATA[Studi Literatur: Prediksi Kata Berikutnya dengan Metode Recurrent Neural Network ]]>
<![CDATA[Trigreisian, Alwizain Almas]]>;
<![CDATA[Harani, Nisa Hanum]]>
<![CDATA[ Jurnal Informatika dan Teknologi Komputer (J-ICOM) Vol 5 No 1 (2024): Jurnal Informatika dan Teknologi Komputer ( JICOM) ]]>
Publisher : <![CDATA[Universitas Samudra ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.55377/j-icom.v5i1.8104
<![CDATA[Next-word prediction is one of the most frequently used tasks in natural language processing. The Recurrent Neural Network (RNN) method is one method that has been proven to be effective in predicting the next word in a sentence, as it is capable of processing text data with order and context. In this research, various algorithms used in the development of next word prediction using the RNN method were analyzed. Some of these algorithms include LSTM (Long Short-Term Memory) and bidirectional LSTM. The results of this research show that the use of the RNN method in predicting the next word is able to provide better results compared to other methods. However, there are still some challenges that need to be overcome in developing the RNN model to predict the next word. Therefore, further research needs to be done in overcoming these challenges so that the use of the RNN method can be further optimized in predicting the next word in a sentence.]]>
<![CDATA[Peningkatan Kinerja Chatbot NLP Asisten: Tinjauan Literatur tentang Metode dan Akurasi dalam Aplikasi Berbasis Percakapan ]]>
<![CDATA[Tri Khaqiqi, M. Ilyas]]>;
<![CDATA[Harani, Nisa Hanum]]>
<![CDATA[ Jurnal Informatika dan Teknologi Komputer (J-ICOM) Vol 5 No 1 (2024): Jurnal Informatika dan Teknologi Komputer ( JICOM) ]]>
Publisher : <![CDATA[Universitas Samudra ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.55377/j-icom.v5i1.8242
<![CDATA[Chatbots have been widely used in various industries such as e-commerce, banking, healthcare, and education to improve efficiency and provide 24/7 services to users. In the field of education, NLP chatbot brings the potential to improve soft skills and hard skills through online learning. This research aims to find suitable methods from previous research to be used in the creation of a conversational chatbot for supporting services of an application system. The research method used is Systematic Literature Review, with comprehensive journal search steps using appropriate keyword search strategies. The research results include 20 articles relevant to the topic of chatbot NLP assistants. The various methods identified in the research include machine learning, deep learning, rule-based approaches, and the use of third-party applications such as Dialogflow and IBM Watson. The analysis results show that the Dynamic Memory Network (DMN) method has the best performance with 91% accuracy. DMN combines the advantages of LSTM and Memory Network with a dynamic attention mechanism, allowing the model to focus on the most relevant information in sequential data. Although this study provides interesting findings, further research is needed to deal with the different characteristics and availability of data in various real-world scenarios. Thus, this article highlights the importance of continuously developing NLP chatbot technology for better applications and improved service quality for users. It is hoped that this article can contribute to the development of research related to NLP chatbot assistants in better and more efficient application systems.]]>
<![CDATA[Next Word Prediction for Book Title Search Using Bi-LSTM Algorithm ]]>
<![CDATA[Trigreisian, Alwizain Almas]]>;
<![CDATA[Harani, Nisa Hanum]]>;
<![CDATA[Andarsyah, Roni]]>
<![CDATA[ The Indonesian Journal of Computer Science Vol. 12 No. 3 (2023): The Indonesian Journal of Computer Science ]]>
Publisher : <![CDATA[AI Society & STMIK Indonesia ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.33022/ijcs.v12i3.3233
<![CDATA[Finding a suitable book title is still quite difficult at the moment. We often guess what book title we want, but in reality the book title is often not available. This research aims to overcome these problems by producing an accurate and efficient prediction model in predicting the next words in book title search using a deep learning algorithm, namely Bidirectional Long Short Term Memory (Bi-LSTM). The research stages consist of data collection, data preprocessing, data modeling, evaluation, and implementation. This research uses a dataset of Indonesian book titles obtained from the bukukita.com online bookstore website with 5618 data. The results show that the resulting deep learning model can predict the next words in the book title search with an accuracy of 81.82%. The model is implemented in the form of a web application using the Django framework, Python language, and MySQL database.]]>
<![CDATA[Probability Prediction for Graduate Admission Using CNN-LSTM Hybrid Algorithm ]]>
<![CDATA[Zuhri, Burhanudin]]>;
<![CDATA[Harani, Nisa Hanum]]>;
<![CDATA[Prianto, Cahyo]]>
<![CDATA[ The Indonesian Journal of Computer Science Vol. 12 No. 3 (2023): The Indonesian Journal of Computer Science ]]>
Publisher : <![CDATA[AI Society & STMIK Indonesia ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.33022/ijcs.v12i3.3248
<![CDATA[Currently, the prediction of student admissions still uses conventional machine learning algorithms where there is no algorithm for optimization. This study aims to produce a model that can predict student acceptance of ownership more optimally by using an optimization hybrid learning algorithm, namely the Convolutional Neural Network Long Short Term Memory (CNN-LSTM). This study uses the Microsoft Team Data Science Process method which consists of business understanding, data acquisition & understanding, modeling, and implementation as well as using the acceptance dataset obtained from the kaggle.com website as much as 500 data. The results showed that the CNN-LSTM hybrid learning model could optimize the prediction of students' chances of success in exposure as evidenced by the evaluation results of RMSE of 6.31%, MAE of 4.4%, and R2 of 80.52%. This model is implemented in a website application using the Python language, the Django framework, and the MySQL database.]]>
<![CDATA[Performance Analysis and Development of QnA Chatbot Model Using LSTM in Answering Questions ]]>
<![CDATA[Ilyas Tri Khaqiqi, M]]>;
<![CDATA[Harani, Nisa Hanum]]>;
<![CDATA[Prianto, Cahyo]]>
<![CDATA[ The Indonesian Journal of Computer Science Vol. 12 No. 3 (2023): The Indonesian Journal of Computer Science ]]>
Publisher : <![CDATA[AI Society & STMIK Indonesia ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.33022/ijcs.v12i3.3249
<![CDATA[This research aims to evaluate the performance of a Long Short-Term Memory (LSTM) based chatbot in answering questions (QnA). LSTM is a type of Recurrent Neural Network (RNN) architecture specifically designed to overcome vanishing gradient problems and can store long-term information. The method used is 5-fold cross-validation to train the chatbot model with 15 epochs at each fold using the dataset provided. The results showed variations in model performance at each fold. At the 5th fold, there was a decrease in performance with 84.63% accuracy, 96.36% precision, 64.9% recall, and 69.84% loss value. This finding shows that there is variability in the performance of the QnA chatbot model at each fold. In conclusion, the LSTM chatbot model can provide good answers with high accuracy and precision. Still, performance variations need to be considered in the use of this chatbot.]]>
<![CDATA[Social Media-Based Sentiment Analysis: Electric Vehicle Usage in Indonesia ]]>
<![CDATA[Salsabila, Helmi]]>;
<![CDATA[Habibi, Roni]]>;
<![CDATA[Harani, Nisa Hanum]]>
<![CDATA[ The Indonesian Journal of Computer Science Vol. 12 No. 3 (2023): The Indonesian Journal of Computer Science ]]>
Publisher : <![CDATA[AI Society & STMIK Indonesia ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.33022/ijcs.v12i3.3250
<![CDATA[This research analyzes the sentiment regarding the use of electric vehicles in Indonesia through social media, utilizing over 10,000 data points from Twitter. The results indicate a variation of positive, negative, and neutral sentiments towards electric vehicles on social media. Female users play a significant role in expressing their views and actively participating in discussions related to electric vehicles. The locations with the highest user activity discussing electric vehicles are Indonesia, DKI Jakarta, Makassar, Tangerang, and Karawang. The peak activity was observed in September 2019, suggesting a significant interest in electric vehicles. The SVM algorithm achieved an accuracy of 88% for positive and neutral sentiments, but performed relatively lower for negative sentiments. This research lacks data on the gender and age of the respondents. Future studies should address these shortcomings to gain a deeper understanding of public perceptions regarding electric vehicles in Indonesia.]]>
<![CDATA[Penerapan Augmented Reality Sebagai Media Promosi Menggunakan Algoritma Regresi Logistik ]]>
<![CDATA[Prianto, Cahyo]]>;
<![CDATA[Harani, Nisa Hanum]]>;
<![CDATA[Andarsyah, Roni]]>
<![CDATA[ Jurasik (Jurnal Riset Sistem Informasi dan Teknik Informatika) Vol 8, No 2 (2023): Edisi Agustus ]]>
Publisher : <![CDATA[STIKOM Tunas Bangsa Pematangsiantar ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.30645/jurasik.v8i2.653
<![CDATA[Nowadays, some companies use social media to promote their product, with no exception PT. Pos Indonesia also promotes hiring an influencer to become a brand ambassador for introducing Pos and its product. In this era, digital marketing has an important influence on business. A unique way is needed to get attention and increase interaction between customers and posted content. For fulfilling that thing, a promotion app with Augmented Reality is designed. This technology combines the virtual and real world at the same time, in Indonesia itself promotion of AR is still seldom. By using AR, the PT Pos promotion package will be shown in the form of 3D objects when the Logo of PT. Pos is highlighted with Augmented Reality Camera. Then the promo could be shared using social media to emerge a bond with the user, so the user will get a poin that is managed by a logistic regression algorithm. Users will feel involved in promotion and also gain benefits in the form of poins so, indirectly there will be a lot of people who promote the product of PT. Pos voluntarily. Modeling using logistic regression is done with 1498 data, 75% of the data becomes the data train and 25% of the rest becomes the data test, the created model has an accuracy 61.07%.]]>
<![CDATA[KINERJA ALGORITMA BACKPROPAGATION DAN RNN DALAM PREDIKSI BEBAN JARINGAN ]]>
<![CDATA[Annisah, Wulan Nur]]>;
<![CDATA[Harani, Nisa Hanum]]>;
<![CDATA[Habibi, Roni]]>
<![CDATA[ Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 7 No 2 (2024) ]]>
Publisher : <![CDATA[Politeknik Bisnis Indonesia ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.37600/tekinkom.v7i2.1586
<![CDATA[Bandwidth allocation optimization is crucial to ensure optimal network performance and user satisfaction. This research aims to identify the best machine learning algorithm between backpropagation and recurrent neural network (RNN) in predicting network load, using two different datasets. The main issue addressed is how to choose the right algorithm for network load prediction to optimize bandwidth allocation. The CRISP-DM methodology was used as the research framework, with four evaluation metrics: Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE). The results showed that the backpropagation algorithm provided the best performance on the data with the lowest evaluation matrix: MAE 0.0203, MSE 0.0007, RMSE 0.0281, and MAPE 20%. In conclusion, the backpropagation algorithm is more suitable for predicting bandwidth requirements compared to RNN based on the evaluation metrics used, making it reliable for bandwidth allocation optimization.]]>
<![CDATA[Predicting Basic Shipping Tariff Using Machine Learning: Prediksi Tarif Dasar Pengiriman Menggunakan Machine Learning ]]>
<![CDATA[Harani, Nisa Hanum]]>;
<![CDATA[Setyawan, M. Yusril Helmi]]>;
<![CDATA[Ferdinan, Dani]]>
<![CDATA[ NUANSA INFORMATIKA Vol. 19 No. 2 (2025): Nuansa Informatika 19.2 Juli 2025 ]]>
Publisher : <![CDATA[FKOM UNIKU ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.25134/ilkom.v19i2.388
<![CDATA[This study explores the application of machine learning algorithms in predicting the Basic Shipping Tariff for logistics, focusing on variables such as Item Price, Shipment Weight, and Distance (KM). Random Forest Regressor and Linear Regression models were used as comparison methods. Experimental results show that the Random Forest Regressor outperforms Linear Regression, achieving an R² value of 0.915 and RMSE of 0.154, while Linear Regression reached an R² value of 0.706 and RMSE of 0.113. Additionally, the Random Forest model achieved lower error values with MSE of 0.000 and MAE of 0.003, compared to Linear Regression with MSE of 0.001 and MAE of 0.007. These error metrics further highlight the superiority of the Random Forest model. In-depth analysis reveals significant relationships between these variables and the Basic Shipping Tariff, showcasing the model's potential application in dynamic pricing strategies within the Indonesian logistics industry. This study aims to contribute to operational efficiency and improve pricing accuracy in the logistics business in Indonesia.]]>
