Garuda - Garba Rujukan Digital

Metode Klasifikasi Kematangan Tandan Buah Segar Kelapa Sawit: Sebuah Tinjauan Sistematis Nurita Evitarina; Kusrini Kusrini
G-Tech: Jurnal Teknologi Terapan Vol 8 No 4 (2024): G-Tech, Vol. 8 No. 4 Oktober 2024
Publisher : Universitas Islam Raden Rahmat, Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70609/gtech.v8i4.5050

Penentuan kematangan buah kelapa sawit sangat penting untuk meningkatkan kualitas dan kuantitas produksi minyak kelapa sawit. Penelitian ini mengkaji penggunaan teknologi deep learning untuk mengklasifikasikan kematangan kelapa sawit melalui Systematic Literature Review (SLR). Metode penelitian yang digunakan adalah Systematic Literature Review (SLR) yang melibatkan analisis 35 jurnal dari Scopus dan Google Scholar dari tahun 2020 hingga 2024, dengan fokus pada kumpulan data, algoritma, lokasi kumpulan data, dan metode pengukuran kinerja model. Hasilnya menunjukkan bahwa ANN dan CNN adalah algoritma yang paling banyak digunakan, dengan penggunaan masing-masing 16% dan 10%. Akurasi, presisi, perolehan, dan skor F1 adalah metrik kinerja yang paling umum. Penelitian di masa depan harus fokus pada peningkatan generalisasi model dan mengintegrasikan data dari berbagai sumber untuk meningkatkan akurasi klasifikasi, tujuannya untuk berkontribusi pada klasifikasi kematangan minyak sawit dan membantu industri meningkatkan efisiensi dan kualitas produksi.

Systematic Literature Review : Klasifikasi Tingkat Kematangan Buah Pisang Suhendri; Kusrini Kusrini
G-Tech: Jurnal Teknologi Terapan Vol 8 No 4 (2024): G-Tech, Vol. 8 No. 4 Oktober 2024
Publisher : Universitas Islam Raden Rahmat, Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70609/gtech.v8i4.5059

Banana ripeness is an important factor that determines the quality, taste and shelf life of the fruit. Manually determining maturity levels tends to be subjective and inconsistent, so a more accurate and efficient automatic system is needed. This research conducted a SLR to evaluate image processing and machine learning techniques in banana ripeness classification CNN is proven to be the most dominant and effective method, with significant accuracy results. Other methods such as kNN, Fuzzy Logic, and ANN also show great potential. The main challenges in developing classification models include image data variability, dataset limitations, and hardware limitations. Recent trends include the use of HSI and multimodal approaches to improve accuracy. Suggestions for future research include collecting larger and more diverse datasets, using data augmentation techniques, exploring HSI sensing, and validating models under real conditions. Thus, this research is expected to make a significant contribution in the development of an automatic system for banana ripeness classification, which can be applied in the agricultural and food industries.

Metode Klasifikasi Tingkat Kematangan Buah dan Sayuran : Tinjauan Sistematis Rama Saktriawindarta; Kusrini Kusrini
G-Tech: Jurnal Teknologi Terapan Vol 8 No 4 (2024): G-Tech, Vol. 8 No. 4 Oktober 2024
Publisher : Universitas Islam Raden Rahmat, Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70609/gtech.v8i4.5067

Maturity classification of fruits and vegetables is important to ensure product quality in the agricultural industry. The aim of this research is to optimize harvest and distribution times using deep learning and machine learning methods. A Systematic Literature Review (SLR) was used to identify effective classification methods and models. The dominant method is image processing (65%), followed by machine learning (50%) and deep learning (42.5%). Models such as CNN, AlexNet, and ResNet-50 show high accuracy. Performance evaluation uses metrics such as accuracy, precision, recall, and F1-score. To improve accuracy, future research is recommended to collect more diverse datasets and use hybrid methods. Development of computing infrastructure and workforce training are also necessary for the application of this technology in the agricultural industry.

Komparasi Algoritma Machine Learning Untuk Menganalisis Sentimen Ulasan Pada Aplikasi Digital Korlantas Polri Siti Delimasari; Kusrini Kusrini
G-Tech: Jurnal Teknologi Terapan Vol 8 No 4 (2024): G-Tech, Vol. 8 No. 4 Oktober 2024
Publisher : Universitas Islam Raden Rahmat, Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70609/gtech.v8i4.5089

Korlantas Polri Digital Application is one of the mobile applications that provides ease for the public in extending the driving license. Sentiment analysis of user reviews can help korlantas polri identify public perception of the given service. The study aims to evaluate which of the five machine learning algorithms performed best from Support Vector Machine (SVM), Naive Bayes, Random Forest, K-Nearest Neighbors (KNN), and Logistic Regression in sentiment analysis. The evaluation was done by measuring accuracy, precision, recall and F1 measure. There were 10,000 reviews labelled with linguistic validation, re-processed, and word weighted after data was collected. Synthetic minority over-sampling techniques (SMOTE) are applied before data splitting for training and testing. The evaluation shows that Random Forest and SVM do the best. Random Forest has an accuracy of 90.77%, recall 90.77%, and its highest F1 rating is 90.79%. SVM has the highest precision with 91.14% among other algorithms, which shows the great potential of both of these algorítms in the analysis of sentiment reviews of digital applications Korlantas Polri.

Metode Machine Learning dan Deep Learning dalam Prediksi Kinerja Siswa: Tinjauan Sistematis Desty Yani; Kusrini Kusrini
G-Tech: Jurnal Teknologi Terapan Vol 9 No 1 (2025): G-Tech, Vol. 9 No. 1 January 2025
Publisher : Universitas Islam Raden Rahmat, Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70609/gtech.v9i1.6501

Research on student performance prediction has advanced rapidly in recent years, driven by the increasing volume of educational data generated by digital learning platforms. This data can be analyzed using Machine Learning (ML) and Deep Learning (DL) techniques, integrated with feature management strategies tailored to specific needs. However, selecting the most relevant features and optimizing predictive models remain significant challenges. Different studies apply various feature selection and engineering techniques, leading to inconsistent results and limited generalizability. This study conducts as a Systematic Literature Review (SLR) to explore ML and DL approaches for student performance prediction, emphasizing their relationship with feature management techniques. The reviewed studies span publications from 2019 to 2024. This SLR aims to assist researchers in identifying effective strategies for predicting student performance, including the selection of methods, datasets, or feature management techniques. Most studies utilized publicly available datasets due to their accessibility and ease of use. Among ML methods, Random Forest emerged as the most frequently applied, achieving an F-measure of 99.5% integration of filter-based and wrapper-based feature selection techniques. Among DL approaches, the ANN-PCACSN model, employing Principal Component Analysis (PCA) for dimensionality reduction, achieved the highest accuracy of 99.32%. These findings highlight the importance of aligning preprocessing strategies with dataset properties and algorithm capabilities to enhance predictive performances.

Application of K-Means and Naïve Bayes Algorithms for Prediction Model of Student Interest Concentration (Case Study: Amikom University Yogyakarta) Danang Eko Prayogo; Kusrini Kusrini
G-Tech: Jurnal Teknologi Terapan Vol 9 No 1 (2025): G-Tech, Vol. 9 No. 1 January 2025
Publisher : Universitas Islam Raden Rahmat, Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70609/gtech.v9i1.6523

Amikom University Yogyakarta, has a Master of Informatics Engineering study program with three concentrations of specialization: Business Intelligence, Digital Information Intelligence, and Intelligence Animation. The choice of concentration by prospective students has been based on subjectivity, not on competence or work experience. To overcome this, this research proposes an algorithm-based concentration prediction and recommendation model to help prospective students choose the appropriate concentration. The dataset is obtained through questionnaires collected from active and inactive students. This research uses the K-Means algorithm for clustering raw data (unsupervised) in order to generate target classes, which are then classified using Naïve Bayes. The clustering process determines concentration labels such as Business Intelligence and others, while the SMOTE technique is used to balance the dataset to avoid data imbalance problems. This approach aims to produce more objective and accurate recommendations in determining student concentrations, reducing the tendency of subjectivity, and increasing the relevance of student competencies to the chosen field of specialization. From this research, the K-Means DBI score is 0.277 and the Naïve Bayes prediction accuracy score is 89%. This research aims to produce more objective and accurate recommendations in determining student concentrations, reducing subjectivity, and increasing the relevance of student competencies to the chosen field of specialization. The proposed model is expected to help universities in designing a more targeted admission strategy, as well as supporting students in making academic decisions that are in accordance with their abilities and interests, thereby increasing the effectiveness of the learning process and the suitability of graduates to the needs of the world of work.

Analysis of Selling Price Determination With Gradient Boosting Algorithm in Traditional Market Stores Bagas Dwi Novianto; Kusrini Kusrini
G-Tech: Jurnal Teknologi Terapan Vol 9 No 3 (2025): G-Tech, Vol. 9 No. 3 July 2025
Publisher : Universitas Islam Raden Rahmat, Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70609/g-tech.v9i3.7264

Traditional market traders often face difficulties in determining optimal selling prices amid competition from modern retailers. This study aims to apply and compare Gradient Boosting and XGBoost algorithms to develop a selling price prediction model for traditional market stores. The research utilizes two datasets : a large-scale dataset from annual sales data and a small-scale dataset from one month of sales. Model training involves hyperparameter optimization using GridSearchCV and evaluation through metrics such as RMSE, MAE, R², and MAPE. Additionally, feature importance and SHAP analyses were conducted to interpret model behavior. The results demonstrate that both models performed well, with R² values nearing 1.0 and MAPE below 2%. Gradient Boosting outperformed XGBoost on the large dataset, while XGBoost showed better accuracy on the small dataset. These findings highlight the potential of machine learning in supporting data-driven pricing strategies for traditional markets.

Title

Found 7 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 7 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 7 Documents
Search