cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Fristi Riandari
Contact Email
hengkitamando26@gmail.com
Phone
+6281381251442
Journal Mail Official
hengkitamando26@gmail.com
Editorial Address
Romeby Lestari Housing Complex Blok C Number C14, North Sumatra, Indonesia
Location
Unknown,
Unknown
INDONESIA
Jurnal Mandiri IT
Published by Institute of Computer Science
ISSN : 23018984     EISSN : 28091884     DOI : https://doi.org/10.35335/mandiri
Core Subject : Science, Education,
Computer Science & IT Library & Information Science Mathematics
The Jurnal Mandiri IT is intended as a publication media to publish articles reporting the results of Computer Science and related research.
Arjuna Subject : Ilmu Komputer (semua kategori) - Kecerdasan Buatan
Articles 187 Documents
 15   16   17   18   19 
Forecasting building permit submissions with fuzzy time series at DPMPTSP Medan Dasopang, Buyung Satrio; Kurniawan, Rakhmat
Jurnal Mandiri IT Vol. 14 No. 1 (2025): July: Computer Science and Field.
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mandiri.v14i1.444

Abstract

Public service is a vital part of government performance, including how the Investment and One-Stop Integrated Services Agency (DPMPTSP) handles building permit applications (IMB). This study aims to estimate the number of IMB applications in Medan City using a method called Fuzzy Time Series (FTS). The forecast is intended as a preliminary step to support better spatial planning, especially as urban building density continues to rise. The FTS method was chosen for its ability to process time series data containing uncertainty. The forecasting process involves several stages: identifying the dataset, setting interval ranges, performing fuzzification, forming fuzzy logical relationships (FLR), grouping fuzzy logical relationship groups (FLRG), applying defuzzification, and measuring accuracy using Mean Absolute Percentage Error (MAPE). The data used include IMB applications from 2022 to 2023, with predictions made for 12 months in 2024. The results show that the FTS model closely follows historical data patterns, evidenced by a MAPE value of 1.99%, which indicates excellent accuracy as it is well below the 10% threshold. A comparative graph between actual and predicted data further supports this, revealing similar trends. In conclusion, the Fuzzy Time Series method is effective for forecasting IMB application volumes and can serve as a valuable reference for urban planning decisions and future time series-based forecasting research involving uncertainty.
Implementation of association method using fp-growth algorithm on sales transaction data at Koperasi Primer Pullahta Hankam Pusdatin KEMHAN RI Aulia, Regifia Ningrum Nur; Prabukusumo, M Azhar; Hidayati, Ajeng
Jurnal Mandiri IT Vol. 14 No. 1 (2025): July: Computer Science and Field.
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mandiri.v14i1.446

Abstract

The conventional recording of sales transaction data frequently results in inaccuracies and presents significant obstacles to comprehensive data analysis. This study was conducted at Primkop Pullahta Hankam Pusdatin Kemhan RI with the aim of generating a product list based on item categories that are most frequently purchased together. These item combinations are expected to assist the cooperative in optimizing sales performance. The research employed a data mining technique known as association rule mining, which is designed to identify and predict customer purchasing behavior through analysis of transaction patterns. The dataset used comprised sales transaction records collected between September and November 2024. The FP-Growth algorithm was selected for its efficiency in identifying frequent itemsets without candidate generation. This algorithm utilized minimum support and confidence thresholds to generate association rules. The modeling process produced five association rules, each meeting the criteria of a minimum support of 20% and a minimum confidence of 80%, indicating strong co-occurrence among specific product combinations. Functional testing using the blackbox method demonstrated that all implemented features performed in accordance with specified functional requirements. The findings offer valuable insights for cooperative management by enabling data-driven decision-making in inventory planning, promotional bundling, and strategic sales targeting. These implications underscore the practical contribution of the research in enhancing operational efficiency and sales strategy within the cooperative sector.
Bangka strait salinity prediction using landsat 9 oli image data Khoirun Nisa; Harsono, Gentio; Martha, Sukendra; Waluyo, Dangan
Jurnal Mandiri IT Vol. 14 No. 1 (2025): July: Computer Science and Field.
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mandiri.v14i1.447

Abstract

Salinity is an important parameter because it affects the environment, such as water quality, growth, and development of aquatic vegetation and various animal species. Conventional water quality monitoring is still ineffective, so it is necessary to utilize technology in monitoring water quality, including water Salinity is an important parameter because it affects the environment, such as water quality, growth, and development of aquatic vegetation and various animal species. Conventional water quality monitoring is still ineffective, so it is necessary to utilize technology in monitoring water quality, including water salinity. Utilization of remote sensing is often used to study salinity both on a small scale and a global scale. Therefore, the author conducted a study to predict salinity in the Bangka Strait using the RRS (Remote Sensing Reflectance) method. The data used are Landsat 9 OLI image data downloaded from the USGS website and in situ salinity data in the Bangka Strait sea. The Landsat 9 OLI image data used is level 2 Surface Reflectance (SR), which is ready for analysis without additional processing by the user. The data obtained were processed using multiple linear regression analysis with Rrs as the independent variable and in situ salinity as the dependent variable. Salinity prediction models are divided into three groups based on the image recording date, namely Rrs 1 for the Landsat 9 OLI image recording on May 9, 2024, Rrs 2 for July 28, 2024, and Rrs 3 for the image recording on September 28, 2023. Multiple linear regression analysis produces R² values for each model of 0.81662874, 0.8170285, and 0.8136894. These R² results indicate that the three models, Rrs 1, Rrs 2, and Rrs 3, are included in the very good criteria in predicting salinity. To choose the best of the three models, by considering the results of the validity test. The NMAE validity test for Rrs 1, Rrs 2, and Rrs 3 is 10.10152, 10.37618, and 8.88680. Meanwhile, the RMSE values are 2.41327, 2.41064, and 2.43253. Therefore, it can be determined that the Rrs 2 model is the best in predicting salinity because it has the highest R² value, namely 0.8170285, and the smallest RMSE, namely 2.41064.
Sentiment analysis of tourist reviews on google maps for pura besakih using machine learning algorithms Sudiatmika, I Putu Gede Abdi; Saputra, Putu Satya; Rahardian, Rifky Lana; Dewi, Komang hari Santhi
Jurnal Mandiri IT Vol. 14 No. 1 (2025): July: Computer Science and Field.
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mandiri.v14i1.449

Abstract

Tourist reviews on digital platforms have become a valuable source of information for understanding visitor experiences. This study applies sentiment analysis to 2,891 Google Maps reviews of Pura Besakih, Bali’s largest and most sacred temple, collected between January 2023 and December 2024. The aim is to assess overall visitor sentiment and identify factors influencing satisfaction and dissatisfaction. Reviews were preprocessed using a standardized pipeline that included translation, cleaning, tokenization, stopword removal, and stemming. Sentiment labeling was conducted using the Indonesian Sentiment Lexicon (InSet), followed by classification using six machine learning models: Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Naive Bayes, Decision Tree, Random Forest, and Neural Network. The SVM model achieved the highest performance with an accuracy of 76.3% and F1-score of 55.68%. Thematic analysis revealed positive feedback highlighting the temple’s spiritual ambiance, architecture, and improved facilities, while negative sentiment was driven by issues such as unauthorized guides, misleading charges, and restricted access. These findings offer valuable insights for tourism stakeholders to improve visitor experience and support sustainable heritage tourism through data-driven decision-making.
Performance analysis of MobileNetV2 based automatic waste classification using transfer learning Firnando, Ricy; Buchari, Muhammad Ali; Marjusalinah, Anna Dwi; Willy; Abdurahman; Isnanto, Rahmat Fadli
Jurnal Mandiri IT Vol. 14 No. 1 (2025): July: Computer Science and Field.
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mandiri.v14i1.451

Abstract

The significant increase in global waste requires innovative and accessible solutions, which aligns with Sustainable Development Goal (SDG) 12, which focuses on reducing the environmental impact of human activities. Automatic waste sorting using Computer Vision and Deep Learning offers a promising alternative to labor-intensive and risky manual methods. This study presents the design, implementation, and comprehensive performance analysis of an automated waste classification system, with a specific focus on evaluating its feasibility on hardware without specialized GPU accelerators. By leveraging transfer learning on a lightweight Convolutional Neural Network (CNN) architecture, MobileNetV2, a model was trained to classify six common waste categories: cardboard, glass, metal, paper, plastic, and other waste. The public “Garbage Classification” dataset from Kaggle, consisting of 2,527 images, was used as the basis for training and validation. The experiment was conducted using the tensorflow-cpu library, which does not require a dedicated GPU accelerator. After 10 training epochs, the model achieved a significant validation accuracy of 86.73%. Computational performance analysis showed an efficient average training time of 31.17 seconds per epoch and a fast average inference time of 14.47 milliseconds per image (~69 FPS) on the validation dataset. These findings demonstrate the feasibility of developing an effective AI-based waste classification system on hardware without a GPU accelerator, providing a realistic performance benchmark for the development of low-cost smart bins with embedded waste sorting solutions in the future, thereby contributing to sustainable waste management practices.
Sentiment analysis of privacy issues in the digital era using the naïve bayes method Ramadhan, Rio Fadli; Kurniawan, Rakhmat
Jurnal Mandiri IT Vol. 14 No. 2 (2025): Computer Science and Field
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mandiri.v14i2.450

Abstract

The development of information technology has triggered public concern about data privacy issues, especially on social media such as X (formerly Twitter). The rampant leaks of personal data have driven the need for a deeper understanding of public opinion. This study aims to analyze public sentiment towards data privacy issues by applying the Naïve Bayes algorithm. The formulation of the problem includes how the public perceives data privacy, how the algorithm performs in classifying sentiment, and how the evaluation results of the model used are. This study uses a quantitative method with a text mining and machine learning approach. Data were taken through crawling techniques on 1,500 tweets related to data privacy. The pre-processing stages were carried out through cleaning, tokenizing, normalization, stopword removal, and stemming. Furthermore, the data was labeled using the InsetLexicon dictionary and weighted using the TF-IDF method. The classification model was built using the Naïve Bayes algorithm and evaluated using accuracy, precision, recall, and f1-score metrics. The results showed that the majority of public opinion on data privacy issues was negative, reflecting concerns over the weak protection of personal data. The Naïve Bayes model performed quite well in sentiment classification. This research is useful in providing insight to the government and digital service providers in developing data protection policies that are more responsive to public opinion.
Reinforcement learning for bitcoin trading: A comparative study of PPO and DQN Prasetyo, Romadhan Edy; Sumanto, Sumanto; Chaidir, Indra; Supriyatna, Adi
Jurnal Mandiri IT Vol. 14 No. 2 (2025): Computer Science and Field
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mandiri.v14i2.455

Abstract

Bitcoin’s high volatility demands automated strategies that adapt to changing market regimes while managing risk. This study compares Proximal Policy Optimization (PPO) and Deep Q-Network (DQN) for Bitcoin trading using hourly BTC/USDT data from 2019 to early 2025. The models are trained to generate buy and sell signals from technical indicators including the Relative Strength Index (RSI), MA20, volatility, Moving Average Convergence Divergence (MACD), volume trend, SMA200, and a weekly trend filter. All features are computed on hourly bars. The evaluation shows that PPO tends to trade more aggressively and delivers higher performance during bullish phases, though with greater risk in unstable markets. By contrast, DQN trades more selectively and maintains better stability in sideways or choppy conditions. These findings support the effectiveness of reinforcement learning for adaptive cryptocurrency trading and highlight complementary strengths between PPO and DQN across market regimes.

Page 19 of 19 | Total Record : 187

 15   16   17   18   19 

Filter by Year

2020 2025


Reset
Filter By Issues
All Issue Vol. 14 No. 2 (2025): Computer Science and Field Vol. 14 No. 1 (2025): July: Computer Science and Field. Vol. 13 No. 4 (2025): April: Computer Science and Field. Vol. 13 No. 3 (2025): January: Computer Science and Field. Vol. 13 No. 2 (2024): CALL FOR PAPER !!! Scope Computer Science Vol. 13 No. 1 (2024): July: Computer Science and Field Vol. 12 No. 4 (2024): April: Computer Science and Field. Vol. 12 No. 3 (2024): January: Computer Science and Field. Vol. 12 No. 2 (2023): October: Computer Science and Field. Vol. 12 No. 1 (2023): July: Computer Science and Field. Vol. 11 No. 4 (2023): April : Computer Science and Field. Vol. 11 No. 3 (2023): January: Computer Science and Field. Vol. 11 No. 2 (2022): October: Computer Science and Field. Vol. 11 No. 1 (2022): July: Computer Science and Field. Vol. 10 No. 2 (2022): January: Computer Science and Field. Vol. 10 No. 1 (2021): July: Computer Science and Field. Vol. 9 No. 2 (2021): January: Computer Science and Field. Vol. 9 No. 1 (2020): July: Computer Science and related. Vol. 8 No. 2 (2020): January: Coputer Science and related. More Issue