Claim Missing Document
Check
Articles

Found 15 Documents
Search

Komparasi Kinerja Algoritma Random Forest dan C4.5 untuk Klasifikasi Harga Mobil Ernawati, Andi; Karim, Abdul
Buletin Ilmiah Informatika Teknologi Vol. 3 No. 1: September 2024
Publisher : AMIK STIEKOM SUMATERA UTARA

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.58369/biit.v3i1.95

Abstract

Determining car prices is a crucial aspect of the automotive industry that requires accurate data analysis for strategic decision-making. This study aims to compare the performance of the Random Forest and C4.5 algorithms in classifying car prices based on specific features, such as technical specifications, production year, and market conditions. The dataset used in this study consists of [mention the size and source of the dataset if available], analyzed using a cross-validation approach to ensure the accuracy of the results. The performance of both algorithms is evaluated based on several metrics, including accuracy, precision, recall, and F1-score. The results show that the Random Forest algorithm consistently outperforms the C4.5 algorithm across most evaluation metrics, achieving an accuracy of [best Random Forest accuracy] compared to [best C4.5 accuracy]. These findings indicate that the Random Forest algorithm is more effective in handling multivariate data complexity and providing more reliable predictions. The conclusions of this study highlight the potential of Random Forest as the primary method for car price classification, especially in scenarios requiring high accuracy levels. This research also contributes to a comparative understanding of decision-tree-based algorithms for applications in the automotive industry and opens opportunities for further research into developing more adaptive and efficient models.
Analisis Perbandingan Algoritma Klasifikasi Data Mining untuk Penentuan Lokasi Perumahan Ernawati, Andi; Iqbal, Muhammad
Journal of Information System Research (JOSH) Vol 6 No 2 (2025): Januari 2025
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/josh.v6i2.6516

Abstract

This study aims to analyze the application of C5.0 and K-Nearest Neighbor (K-NN) algorithms in the classification process for determining the optimal location for housing. The classification process involves several factors such as land price, accessibility, public facilities, crime rate, infrastructure, land availability, and consumer preferences. The research conducted tests on both algorithms to compare their performance in generating accurate predictions. The results show that the C5.0 algorithm outperforms K-NN, achieving an accuracy rate of 100%, compared to K-NN, which achieved an accuracy of 66.67%. This demonstrates that C5.0 is more effective in modeling data and producing more precise classifications. Therefore, it can be concluded that the use of data mining algorithms, particularly C5.0, greatly assists in the classification process for determining housing locations, providing more optimal results compared to K-NN.
Penerapan Data Mining Untuk Klasifikasi Penduduk Miskin Di Kabupaten Labuhanbatu Menggunakan Random Forest Dan K-Nearest Neighbors Ernawati, Andi; Khairul; Sitorus, Zulham; Iqbal, Muhammad; Nasution, Darmeli
Bulletin of Information Technology (BIT) Vol 6 No 2: Juni 2025
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bit.v6i1.1783

Abstract

This study aims to apply and compare the performance of two data mining algorithms—Random Forest (RF) and K-Nearest Neighbors (KNN)—in classifying poverty status among residents of Labuhanbatu Regency. The dataset includes information on occupation, income, housing, and education from 21,137 individuals. After undergoing preprocessing, model training, hyperparameter optimization, and evaluation, both models were assessed using five key metrics: accuracy, precision, recall, F1-score, and AUC. The results show that Random Forest performed slightly better than KNN, achieving an accuracy of 0.6023, precision of 0.4827, recall of 0.4177, F1-score of 0.4479, and an AUC of 0.5681. In comparison, KNN obtained an accuracy of 0.5990, precision of 0.4771, recall of 0.4006, F1-score of 0.4355, and an AUC of 0.5622. Based on these findings, it can be concluded that Random Forest is more effective for poverty classification on this dataset, although the performance difference is relatively small.
Uncovering Smartphone Brand Strategies through Specification-Based Clustering and Classification Karim, Abdul; Ernawati, Andi
Buletin Ilmiah Informatika Teknologi Vol. 4 No. 1: September 2025
Publisher : AMIK STIEKOM SUMATERA UTARA

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.58369/biit.v4i1.167

Abstract

In an increasingly saturated smartphone market, brand differentiation through technical specifications has become a core strategy for attracting diverse consumer segments. This study proposes a machine learning approach to uncover underlying brand strategies by leveraging smartphone specifications and market pricing across multiple regions. We utilize unsupervised clustering algorithms (K-Means, DBSCAN) to segment devices based on technical features, followed by supervised classification models (Random Forest, XGBoost) to identify and interpret brand-driven design strategies. The dataset comprises smartphones released in 2024–2025, including attributes such as RAM, camera specifications, processor type, battery capacity, and launch prices in Pakistan, India, China, USA, and Dubai. Our findings reveal distinct clusters that align with different pricing tiers and show clear brand positioning patterns. Feature importance analysis using SHAP highlights battery capacity, screen size, and processor type as critical variables influencing brand classification. This study provides valuable insights for both manufacturers and consumers in understanding competitive product strategies within the global smartphone market.
EFEKTIVITAS ANTIBAKTERI EKSTRAK DAUN TEH HIJAU TERHADAP BAKTERI MYCOBACTERIUM TUBERCULOSIS Misnarliah; Ernawati, Andi
Jurnal Biogenerasi Vol. 10 No. 1 (2024): Volume 10 Nomor 1, Agustus 2024 - Februari 2025
Publisher : Universitas Cokroaminoto Palopo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30605/biogenerasi.v10i1.4431

Abstract

Tuberculosis is still a global health problem. Prolonged treatment of tuberculosis and using several anti-tuberculosis drugs (OAT) can cause side effects, one of which is multidrug resistance. Cases of multidrug-resistant (MDR) and extensively drug-resistant (XDR) Mycobacterium strains continue to increase. . Research and development of active compounds from medicinal plants to achieve more effective tuberculosis treatment is still being promoted. This research aims to determine the potential and effectiveness of green tea plant (Camellia sinensis) leaf extract in inhibiting the growth of Mycobacterium tuberculosis strain H37Rv bacteria using the LJ (Lowenstein-Jensen) method in vitro. The research samples used were green tea plant leaves (Camellia sinensis) obtained from the Malino Village Tea Plantation, Tinggi Moncong Regency, Gowa District, South Sulawesi. The leaf extract was made in 4 types of concentrations, namely 10, 20, 50 and 100 µg/ml, each of which was tested against the clinical isolate of Mycobacterium tuberculosis strain H37Rv using the LJ (Lowenstein-Jensen) method as the standard for tuberculosis examination. Of the four extracts tested in vitro, only extract concentrations of 50 and 100 µg/ml were able to very strongly inhibit and kill the growth of Mycobacterium tuberculosis strain H37Rv (inhibition percentage of 100%), not a single bacterial colony growth was found during the observation period. The percentage of inhibition of green tea plant leaves (Camellia sinensis) is the same as the percentage of inhibition of the drug rifampicin. Thus, the leaves of the green tea plant (Camellia sinensis) with concentrations of 50 and 100 µg/ml have potential antituberculosis activity and are prospective to be developed as antituberculosis from natural ingredients, and also as an additional therapeutic complement for TB.