Jurnal Buana Informatika

Implementasi Algoritma Apriori sebagai Association Rule Learning untuk Mengidentifikasi Pola Item Dataset Penjualan

2025-04-15T18:58:30+07:00

Retail store competition is becoming more intense, so marketing and product arrangement are crucial for shopping efficiency, maintaining comfort, and increasing profits. This study analyzes consumer shopping habits for goods in each transaction through market basket analysis. The Apriori algorithm is a common technique for finding frequent item search techniques in building association rules, namely the relationships between item combinations in a dataset. The aim is to implement the Apriori algorithm as an association rule learning method to identify patterns within sales data. The Apriori association rule is compared to the frequent pattern growth algorithm, which finds the most frequently occurring patterns in a dataset. Based on the tests, the average lift ratio for the Apriori algorithm is 1.58, while for the frequent pattern growth algorithm, it is 1.28. This indicates that the Apriori algorithm performs better than the frequent pattern growth algorithm.

Sistem Pakan Cerdas Berbasis IoT Untuk Optimalisasi Peternakan Kambing Umbaran di Era Digital Farm

2025-04-12T14:12:53+07:00

This research aims to create a smart feeding system based on the Internet of Things (IoT) to enhance the efficiency of feed delivery in goat farming. The system automatically regulates the feed dispenser according to a predetermined schedule, making it easier for farmers to manage feed. System testing demonstrates its effectiveness in reducing feed delivery time and minimizing waste. The system features an LCD screen that displays the dispenser status, providing real-time information to farmers. This technology also allows for remote monitoring, enabling farmers to manage feed more effectively. The implementation of this system is expected to improve productivity and animal welfare while promoting modernization in farming practices in Indonesia. This innovation is anticipated to offer a sustainable solution to challenges in feed management, providing long-term benefits for farmers and the livestock industry.

Importance of Feature Selection for Multiple Disease Classification

2025-04-11T19:42:52+07:00

The performance of machine learning in disease classification heavily depends on effective feature selection. This study explores feature selection methods—Boruta and Recursive Feature Elimination (RFE)—with ensemble models like Random Forest, Decision Tree, Gradient Boosting, LightGBM, and XGBoost using Electronic Health Records (EHR) data. Results show that combining Boruta with LightGBM achieves the highest accuracy of 99%. Feature selection enhances precision by focusing on relevant variables and removing unnecessary ones. Further analysis reveals that features such as Red Blood Cells, Insulin, Heart Rate, and Cholesterol significantly influence the classification of specific diseases. These findings highlight the importance of feature selection in multi-disease classification and medical data analysis, improving the efficiency of machine learning systems. Future research should develop more flexible feature selection methods and test models on diverse disease datasets.

Ekstraksi Pengetahuan dari Ulasan Aplikasi CapCut Menggunakan Metode Aspect-Based Sentiment Analysis dan Klasifikasi

2025-05-05T11:21:43+07:00

Indonesia is experiencing rapid technological development, especially in the use of the internet and editing platforms like CapCut. These platforms enable video editing on various devices; however, user satisfaction is not always guaranteed due to individual differences in experience. This research aims to identify user sentiment towards the CapCut application based on aspects, using an Aspect-Based Sentiment Analysis (ABSA) approach supported by Machine Learning algorithms for the aspect-based sentiment classification task. The algorithm used in the classification process is Support Vector Machine. The data used are reviews of the CapCut application from the Google Play Store, with a total of 22,668 data points. The results show that the Support Vector Machine (SVM) algorithm performs well in each aspect, with accuracy values of 0.88 for the feature aspect and 0.87 for the user experience aspect. The results of knowledge extraction are obtained in the form of XML, which contains user sentiment information on two main aspects: features and user experience.

Peningkatan Akurasi Rekomendasi Dokter pada Kondisi Data Sparsity Menggunakan Algoritma Content-Based Filtering

2025-01-31T20:37:17+07:00

The growth of healthcare applications such as Halodoc, Alodokter, and Klikdokter has enabled easier access to doctor recommendations. However, generating relevant recommendations remains challenging. One key issue is data sparsity, where limited doctor attributes reduce the system’s accuracy. This study develops a doctor recommendation system using a Content-Based Filtering (CBF) approach based on five main attributes: specialization, rating, consultation fee, years of practice, and gender. Data imputation and attribute weighting techniques are applied to enhance accuracy. Results show that the proposed method reduces the Mean Absolute Error (MAE) from 0.142 to 0.102 and the Root Mean Squared Error (RMSE) from 0.205 to 0.150. These findings indicate that the implemented techniques improve the recommendation system under sparse data conditions.