Article Per Year (5 Year)
p-Index From 2020 - 2025
1.829
P-Index
This Author published in this journals
All Journal Media Statistika Mechatronics, Electrical Power, and Vehicular Technology Jurnal INKOM Jurnal Teknologi Informasi dan Ilmu Komputer Jurnal Pilar Nusa Mandiri Jurnal Linguistik Komputasional Journal of Applied Data Sciences Jurnal Pengabdian Ibnu Sina
Pardede, Hilman Ferdinandus
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Education Electrical & Electronics Engineering Engineering Environmental Science Health Professions Social Sciences

Published : 11 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 11 Documents
Search

 1   2 
Teknik Normalisasi Fitur Secara Adaptif untuk Sistem Pengenalan Ucapan Tahan Terhadap Gema Pardede, Hilman Ferdinandus
INKOM Journal Vol 10, No 2 (2016)
Publisher : Pusat Penelitian Informatika - LIPI

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (290.672 KB) | DOI: 10.14203/j.inkom.475

Abstract

Gema menurunkan performa sistem pengenalan ucapan (SPU) atau automatic speech recognition secara signifikan. Salah satu teknik yang paling populer untuk mengurangi efek gema adalah dengan menormalisasi fitur pada SPU. Pada penelitian sebelumnya, q-log spectral mean normalization (q-LSMN) telah diperkenalkan untuk mengurangi efek distorsi aditif dan convolutif. Metode ini merupakan pengembangan teknik normalisasi konvensional pada domain q-log. Metode inidikembangkan untuk mengurangi efek gema dan teknik adaptif untuk menentukan nilai q terbaik untuk q-LSMN diperkenalkan. Hasil percobaan pada pengenalan angka (digit recognition) menunjukkan bahwa teknik tersebut meningkatkan ketahanan SPU terhadap gema. Metode ini lebih baik dibandingkan metode normalisasi konvensional seperti cepstral mean normalization dan log spectral mean normalization.
Robust remaining useful life prediction of lithium-ion battery with convolutional denoising autoencoder Yuliani, Asri Rizki; Pardede, Hilman Ferdinandus; Ramdan, Ade; Zilvan, Vicky; Yuwana, Raden Sandra; Amri, M Faizal; Kusumo, R. Budiarianto Suryo; Pramanik, Subrata
Journal of Mechatronics, Electrical Power, and Vehicular Technology Vol 15, No 1 (2024)
Publisher : National Research and Innovation Agency

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55981/j.mev.2024.905

Abstract

Using lithium-ion (Li-ion) batteries exceeding their useful lifetime may be dangerous for users, and hence, developing an accurate prediction system for batteries that remain useful for life is necessary. Many deep learning models, such as gated recurrent units and long short-term memory (LSTM), have been proposed for that purpose and have shown good results. However, their performance when dealing with noisy data degrades significantly. This may hamper their implementations for the real world since battery data are prone to noise. In this paper, we develop a robust prediction model in a noisy environment for predicting the remaining useful life (RUL) of Li-ion batteries. We propose a denoising autoencoder (DAE) utilized to remove noise from the data. The DAE is built with convolutional layers instead of traditional feed-forward networks here. We combine DAE with LSTM as the predictor. The proposed framework is evaluated using artificially corrupted battery data provided by National Aeronautics and Space Administration (NASA). The results reveal that our proposed method improves robustness when data contain various types of noise. A comparative study using the traditional approach has also been conducted. Our evaluation shows that convolutional layers are more effective than the traditional approach and that the original composition of the DAE was built using traditional feed-forward networks. DAE with convolutional layers has the best average performance with MSE of 0.61 and is the most consistent model.
Distracted driver behavior recognition using modified capsule networks Kadar, Jimmy Abdel; Dewi, Margareta Aprilia Kusuma; Suryawati, Endang; Heryana, Ana; Zilfan, Vicky; Kusumo, Budiarianto Suryo; Yuwana, Raden Sandra; Supianto, Ahmad Afif; Pratiwi, Hasih; Pardede, Hilman Ferdinandus
Journal of Mechatronics, Electrical Power, and Vehicular Technology Vol 14, No 2 (2023)
Publisher : National Research and Innovation Agency

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14203/j.mev.2023.v14.177-185

Abstract

Human activity recognition (HAR) is an increasingly active study field within the computer vision community. In HAR, driver behavior can be detected to ensure safe travel. Detect driver behaviors using a capsule network with leave-one-subject-out validation. The study was done using CapsNet with leave-one-subject-out validation to identify driving habits. The proposed method in this study consists of two parts, namely encoder and decoder. The encoder used in this study modifies Sabour’s capsule network architecture by adding a convolution layer before going to the primary capsule layer. The proposed method is evaluated using a primary dataset with 10 classes and 300 images for each class. The dataset is split based on hold-out validation and leave-one-subject-out validation. The resulting models were then compared to conventional CNN architecture. The objective of the research is to identify driving behavior. In this study, the proposed method results an accuracy rate of 97.83 % in the split dataset using hold-out validation. However, the accuracy decreased by 53.11 % when the proposed method was used on a split dataset using leave-one-subject-out validation. This is because the proposed method extracts all features including the attributes of each participant contained in the input image (user-independent). Thus, the resulting model in this study tends to overfit.
Analisis Komentar Cyberbullying Terhadap Kata Yang Mengandung Toksisitas Dan Agresi Menggunakan Bag of Words dan TF-IDF Dengan Klasifikasi SVM Krisnandi, Dwi; Ambarwati, Rini Novi; Asih, Anggun Yuli; Ardiansyah, Angga; Pardede, Hilman Ferdinandus
Jurnal Linguistik Komputasional Vol 6 No 2 (2023): Vol. 6, No. 2
Publisher : Indonesia Association of Computational Linguistics (INACL)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26418/jlk.v6i2.85

Abstract

Masyarakat menjadi orang yang anti sosial tidak banyak berkomunikasi dengan lingkungan dan mereka lebih menyukai komunikasi melalui internet atau social media. dengan internet atau social media orang bisa menyembunyikan identitas aslinya dan bisa mengakibatkan keburukan saat berkomunikasi atau memberi sebuah komentar di social media dengan menggunakan kata-kata yang buruk, menghina, tidak sopan mempermalukan, mengancam, mengganggu menghina, mengucilkan, merusak reputasi orang melalui internet atau teknologi digital. Hal seperti ini termasuk dalam tindak kejahatan yang dinamakan cyberbullying. Cyberbullying berdampak buruk bagi korbannya seperti mengakibatkan depresi, hingga yang terburuk hingga mengakibatkan seseorang melakukan bunuh diri. Cyberbullying sering menjadi masalah besar baik di tingkat nasional maupun global. Pada penilitian ini akan dibahas mengenai text mining alisis sentiment cyberbullying dengan menentukan kata yang mengandung toxicity dan aggression. Metode yang digunakan pada penelitian ini memakai model Bag of Word dan TF-IDF dengan klasifikasi SVM. Hasil dari penelitian ini didapat akurasi tertinggi dengan menggunakan model Bag of Word dengan akurasi sebesar 65,2% di banding dengan pemodelan menggunakan TF-IDF dengan akurasi sebesar 64,7% Hal ini menunjukkan bahwa penerapan Bag of Word mampu menghasilkan tingkat akurasi yang lebih baik dalam memprediksi kata yang mengandung cyberbullying dari penelitian ini dibandingkan dengan model TF-IDF.
Implementation of Scale-Invariant Feature Transform Convolutional Neural Network for Detecting Distracted Driver Fhadilla, Nahdatul; Sulandari, Winita; Susanto, Irwan; Slamet, Isnandar; Sugiyanto, Sugiyanto; Subanti, Sri; Zukhronah, Etik; Pardede, Hilman Ferdinandus; Kadar, Jimmy Abdel
Journal of Applied Data Sciences Vol 5, No 3: SEPTEMBER 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i3.222

Abstract

A distraction while driving a vehicle may result in fatal consequences, namely accidents that may leave road users seriously injured or even dead. In order to mitigate this risk, it is imperative to establish a distracted driver detection system that is both precise and real-time. This research focuses on the application of artificial intelligence, with a particular emphasis on deep learning, which is achieved through the utilization of the Convolutional Neural Network (CNN) model. In order to enhance the detection of inattentive drivers and produce a more precise model, a scaleinvariant feature transform (SIFT)-CNN combination is proposed. The activities of the driver while operating a vehicle are categorized into ten categories in this study. One of these categories is considered a normal condition, while the remaining nine are classified as inattentive behaviors. This study implemented Adam optimization with 64 batches, a learning rate of 0.001, and epochs of 20, 25, 50, and 100. The proposed CNNSIFT model is capable of achieving superior performance in comparison to the solitary CNN model, as evidenced by the experimental results. The CNN-SIFT model has achieved 99% accuracy and a 0.05 loss when the hyperparameter configuration is optimized for 50 epochs. The analysis indicates that the accuracy of the features obtained from CNN-SIFT can be improved by approximately 1% compared with CNN to classify the type of driver distraction behavior. The model's reliability was further enhanced by its evaluation on test data, which resulted in high accuracy, precision, recall, and F1-score values. The model's ability to accurately identify driver behavior with a high degree of reliability is demonstrated by these results, which are a positive contribution to the improvement of road safety.
IMPLEMENTATION OF PROPHET IN AMERICAN ELECTRICITY FORECASTING WITH AND WITHOUT PARAMETER TUNING Sulandari, Winita; Yudhanto, Yudho; Hapsari, Riskhia; Wijayanti, Monica Dini; Pardede, Hilman Ferdinandus
MEDIA STATISTIKA Vol 17, No 1 (2024): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.17.1.93-104

Abstract

Prophet is one of the machine learning approximation methods that accommodate trends, seasonality, and holiday impacts in time series data. Generally, the performance of machine learning models can be improved by implementing hyperparameter tuning. This study investigates whether hyperparameter tuning can improve the model's performance. To show its effectiveness, the Prophet model constructed by parameter tuning is compared to the one with fixed parameter values (namely the default model) for both the original series and the Box-Cox transformed series in terms of mean absolute percentage error (MAPE). Based on the experimental results of the twenty-four daily electricity load time series in American Electric Power (AEP). This shows that parameter tuning successfully reduces the MAPE of the default model in the range of about 3-8% for training data. However, there is no guarantee for testing data. Although, in some cases, parameter tuning can reduce the MAPE value of the default model by up to 38%, in other cases, it actually increases the MAPE of the default model by almost 15%. The experiments on testing data also show that models built from transformed data do not necessarily produce more accurate forecast values than those built from the original data.
SOSIALISASI KESADARAN KEAMANAN SIBER PADA BADAN SANTUNAN YATIM KELURAHAN PONDOK CINA DEPOK Riana, Dwiza; Ernawan, Ferda; Na'am, Jufriadif; Pardede, Hilman Ferdinandus; Hasanah, Riyan Latifahul
Jurnal Pengabdian Ibnu Sina Vol. 4 No. 1 (2025): Januari 2025
Publisher : LPPM Universitas Ibnu Sina

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36352/j-pis.v4i1.848

Abstract

ABSTRAK Perkembangan teknologi informasi dan komunikasi selain membawa manfaat tapi juga berdampak dengan munculnya tindak pidana baru dalam bidang teknologi. Modus kejahatan dalam bidang teknologi (cyber chrime) mengikuti alur yang terjadi dalam sistem digital. Lima jenis penipuan yang sering terjadi yaitu, penipuan berkedok hadiah, pinjaman digital ilegal, pengiriman tautan yang berisi malware atau virus, penipuan berkedok krisis keluarga, dan investasi ilegal. Dalam rangka melaksanakan kegiatan Tri Dharma Perguruan Tinggi yaitu Pengabdian kepada Masyarakat, Fakultas Teknologi Informasi Universitas Nusa Mandiri menyelenggarakan pelatihan dengan tema “Sosialisasi Kesadaran Keamanan Siber pada Badan Santunan Yatim Kelurahan Pondok Cina Depok”. Kegiatan ini bertujuan untuk meningkatkan kesadaran peserta akan pentingnya keamanan siber serta menambah wawasan mengenai keamanan siber. Luaran pengabdian kepada masyarakat berupa press release yang ditayangkan di Nusa Mandiri News serta artikel jurnal. Kata Kunci: keamanan siber, pengabdian kepada masyarakat, sosialisasi ABSTRACT The development of information and communication technology not only brings benefits but also has an impact on the emergence of new criminal acts in the field of technology. The mode of crime in the field of technology (cyber crime) follows the flow that occurs in digital systems. The five types of fraud that often occur are, fraud under the guise of gifts, illegal digital loans, sending links containing malware or viruses, fraud under the guise of family crises, and illegal investments. In order to carry out the Tri Dharma of Higher Education activities, namely Community Service, the Faculty of Information Technology, Nusa Mandiri University held training with the theme "Socialization of Cyber ​​Security Awareness in the Orphan Compensation Agency, Pondok Cina Village, Depok". This activity aims to increase participants' awareness of the importance of cyber security and increase their insight into cyber security. The output of community service is in the form of press releases published in Nusa Mandiri News and journal articles. Keywords: cyber security, community service, socialization
MACHINE LEARNING FOR EMPLOYMENT POSITION MAPPING Apriadi, Sena Aditia; Pardede, Hilman Ferdinandus
Jurnal Pilar Nusa Mandiri Vol. 21 No. 2 (2025): Pilar Nusa Mandiri : Journal of Computing and Information System Publishing Pe
Publisher : LPPM Universitas Nusa Mandiri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33480/pilar.v21i2.3028

Abstract

Employee performance directly impacts organizational efficiency, yet traditional HR analytics often lack predictive precision. This study bridges HR theory and machine learning by evaluating tree-based algorithms for employee data analysis. Using a dataset of 15,227 employee records, we tested the Bagged Decision Tree algorithm, focusing on variables such as talent, career values, and aspirations. The Bagged Decision Tree achieved 98.65% accuracy, with talent and career values as key predictors. Excluding aspiration values reduced accuracy slightly to 98.57%, while excluding career values lowered it significantly to 92.13%. These findings highlight the robustness of the Bagged Decision Tree in HR analytics and emphasize the importance of variable selection, particularly career values and talent, in predicting performance outcomes. Future work should further explore real-world implementation challenges.
Autoencoder untuk Sistem Prediksi Berat Lahir Bayi Nugraha, Fitra Septia; Pardede, Hilman Ferdinandus
Jurnal Teknologi Informasi dan Ilmu Komputer Vol 9 No 2: April 2022
Publisher : Fakultas Ilmu Komputer, Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25126/jtiik.2022923868

Abstract

Salah satu ukuran terpenting saat awal persalinan adalah keakuratan prediksi berat lahir. Dengan menggunakan metode prediksi yang tepat, perkiraan ekstrim berat lahir bayi dapat dideteksi lebih atau kurang sehingga beberapa tindakan pencegahan dapat dilakukan sebelum persalinan. Di sisi lain, untuk meningkatkan akurasi suatu proses pembelajaran, dibutuhkan suatu prediksi yang akurat untuk masalah yang dihadapi atau dengan menggunakan gabungan beberapa metode. Penelitian bertujuan untuk prediksi berat lahir bayi menggunakan metode Deep Learning autoencoder untuk memprediksi berat lahir bayi. Salah satu tantangan dalam pengembangan sistem prediksi berat lahir bayi adalah datanya yang berdimensi tinggi. Teknik konvensional untuk feature reduction seperti principal component analysis (PCA), mengasumsikan orthogonality atau independensi antar komponen prinsipal nya. Dengan autoencoder, asumsi tersebut tidak ada. Sehingga autoencoder dapat memodelkan korelasi antar fitur. Dengan melakukan variasi parameter pada autoencoder, performa terbaik diperoleh adalah MSE 0.002, MAE 0.029, R2 0.991 dengan autoencoder dengan 4 lapisan hidden layer encoder dan decoder. Ini lebih baik dibandingkan PCA. AbstractOne of the most important measurements at the onset of labor is the accuracy of the prediction of birth weight. By using precise prediction methods extreme estimates of baby birth weight can be detected more or less so that some precautions can be taken before delivery. On the other hand, to improve the accuracy of a learning process, an accurate prediction is needed for the problem at hand or by using a combination of several methods. This study aims to predict baby birth weight using the Deep Learning autoencoder method to predict baby birth weight. One of the challenges in developing a predictive system for infant birth weight is the high dimensional data. Conventional techniques for feature reduction, such as principal component analysis (PCA), assume orthogonality or independence between the principal components. With an autoencoder, that assumption doesn't exist. So that the autoencoder can model the correlation between features. By varying the parameters of the autoencoder, the best performance is MSE 0.002, MAE 0.029, R2 0.991 with an autoencoder with 4 hidden layer encoder and decoder layers. This is better than PCA.
Peningkatan Performa Pengelompokan Pola Berpikir Siswa dalam Belajar pada Media Pembelajaran Menggunakan Direct Batch Growing Self Organizing Map Izzuddin, Mochammad; Supianto, Ahmad Afif; Tibyani, Tibyani; Pardede, Hilman Ferdinandus; Yuliani, Asri Rizki; Ramdan, Ade
Jurnal Teknologi Informasi dan Ilmu Komputer Vol 9 No 1: Februari 2022
Publisher : Fakultas Ilmu Komputer, Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25126/jtiik.2022915573

Abstract

Tidak sedikit siswa mengalami kendala untuk keluar dari kebuntuan berpikir saat belajar. Setiap siswa memiliki caranya masing-masing untuk menyelesaikan masalah kebuntuan tersebut, yang disebabkan oleh pola berpikir yang berbeda-beda. Kendati berbeda, pola berpikir tersebut memiliki kemiripan yang dapat dikelompokkan agar pemberian umpan balik dapat dilakukan dengan tepat secara berkelompok. Salah satu cara yang dapat dilakukan untuk mengelompokkan siswa berdasarkan pola berpikirnya adalah clustering. Penelitian untuk pengelompokkan berdasarkan kecerdasan sudah pernah dilakukan menggunakan salah satu teknik clustering yaitu Self Organizing Map (SOM). Namun SOM memiliki keterbatasan dalam menentukan ukuran jaringan karena bersifat statis. Keterbatasan yang ada pada SOM dapat diatasi, penelitian ini mengusulkan Direct Batch Growing Self Organizing Map (DBGSOM) yang bersifat dinamis dalam ukuran jaringan dan lebih cepat dalam proses pelatihannya. Penelitian ini dimulai dengan mengidentifikasi masalah untuk mengetahui kemungkinan penyelesaian permasalahan. Tahap selanjutnya adalah pengumpulan data dan pemilihan data yang digunakan dalam penelitian. Tahap akhir, evaluasi dilakukan terhadap data yang terdiri dari 12 assignment untuk mengetahui performa terbaik dari DBGSOM. Hasil evaluasi yang telah dilakukan menunjukkan bahwa clustering DBGSOM memperoleh performa lebih baik daripada SOM pada 11 assignment dari 12 assignment. Pengukuran signifikansi perbandingan dilakukan dengan metode Wilcoxon yang menghasilkan nilai test stat 8 dan critical value 13. Hal ini membuktikan bahwa penerapan DBGSOM mampu memberikan peningkatkan performa clustering yang signifikan dari SOM. AbstractA few times, students have difficulty getting out of the deadlock in thinking while studying. Each student has their own way of solving the deadlock problem, which is caused by different thinking patterns. Although different, these thinking patterns have similarities that can be grouped so that giving feedback can be done appropriately in groups. One way that can be done to group students based on their thinking patterns is clustering. Research for grouping based on intelligence has been done using one of the clustering techniques, namely Self Organizing Map (SOM). However, SOM has limitations in determining network size because it is static. The limitations that exist in SOM can be overcome, this study proposes a Direct Batch Growing Self Organizing Map (DBGSOM) which is dynamic in network size and faster in the training process. This research begins by identifying the problem to determine the possibility of solving the problem. The next stage is data collection and data selection used in research. The final stage, evaluation is carried out on the data consisting of 12 assignments to find out the best performance of DBGSOM. The results of the evaluation that have been carried out show that DBGSOM clustering has better performance than SOM on 11 assignments out of 12 assignments. The comparison significance measurement was carried out using the Wilcoxon method which resulted in a test stat value of 8 and a critical value of 13. This proves that the application of DBGSOM is able to provide a significant increase in clustering performance from SOM.
Co-Authors Ade Ramdan Ahmad Afif Supianto Ambarwati, Rini Novi Amri, M Faizal Angga Ardiansyah Apriadi, Sena Aditia Asih, Anggun Yuli Dewi, Margareta Aprilia Kusuma Dwiza Riana Ferda Ernawan Fhadilla, Nahdatul Hapsari, Riskhia Hasanah, Riyan Latifahul Heryana, Ana Irwan Susanto Isnandar Slamet, Isnandar Izzuddin, Mochammad Kadar, Jimmy Abdel Krisnandi, Dwi Kusumo, Budiarianto Suryo Kusumo, R. Budiarianto Suryo Na'am, Jufriadif Nugraha, Fitra Septia Pramanik, Subrata Prasetyo, Onky pratiwi, hasih Sri Subanti Sugiyanto - Suryawati, Endang Syaiful Anam Tibyani Tibyani, Tibyani Vicky Zilvan, Vicky Wijayanti, Monica Dini Winita Sulandari Yudho Yudhanto, Yudho Yuliani, Asri Rizki Yuwana, Raden Sandra Zilfan, Vicky Zukhronah, Etik