Claim Missing Document
Check
Articles

Found 2 Documents
Search
Journal : JOURNAL OF INFORMATION SYSTEM RESEARCH (JOSH)

Perbandingan Algoritma Naïve Bayes dan K-Nearest Neighbor (K-NN) Untuk Klasifikasi Penyakit Gagal Jantung Zahri, Firman; Insani, Fitri; Jasril, Jasril; Oktavia, Lola
Journal of Information System Research (JOSH) Vol 6 No 2 (2025): January 2025
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/josh.v6i2.6480

Abstract

A condition known as heart failure, where the heart is unable to pump enough blood to meet the body's needs for oxygen and nutrients, should not be taken lightly. This can result in a number of symptoms, such as fatigue, fluid retention, and dyspnea. The World Heart Federation estimates that up to 1.8 million people in Southeast Asia suffered from heart failure in 2014. For prompt and efficient treatment, heart failure is a medical problem that needs to be identified. This disease has the potential to worsen if not treated immediately. Several machine learning methods can be used to help diagnose and categorize this disease. One of them is the popular algorithm, namely Naive Bayes and K-Nearest Neighbors. Naive Bayes is a simple but very efficient probability-based machine learning algorithm, especially in classification applications. K-Nearest Neighbors is comparing the data to be predicted with a number of its closest data in a feature space based on a certain distance, such as Euclidean distance, Manhattan, or others. This study was conducted using Confusion Matrix to evaluate and compare the Naive Bayes and K-Nearest Neighbor algorithms in the categorization of heart failure disease by collecting data totaling 918 heart failure patient data from kaggle. Based on the research findings, the K-Nearest Neighbor method achieved an accuracy score of 76%, while the Naive Bayes approach achieved 90% accuracy using a ratio of 80:20.
Penerapan Algoritma FP-Growth dan K-Means Clustering dalam Analisis Pola Asosiasi Berdasarkan Segmentasi Pelanggan Hasibuan, Aldiansyah Pramudia; Insani, Fitri; Nazir, Alwis; Afrianty, Iis
Journal of Information System Research (JOSH) Vol 6 No 3 (2025): April 2025
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/josh.v6i3.7112

Abstract

The pharmaceutical industry has experienced rapid growth, urging companies to leverage sales data effectively to enhance data-driven marketing strategies. However, utilizing sales data remains a challenge for XYZ company, a pharmaceutical distributor. This study aims to analyze customer purchasing patterns by applying the FP-Growth algorithm for association analysis, combined with customer segmentation using the K-Means algorithm based on RFM (Recency, Frequency, Monetary) analysis. The segmentation process resulted in four customer clusters: active and loyal customers (Cluster 1), passive customers (Cluster 2), less active customers (Cluster 3), and new customers (Cluster 4). FP-Growth analysis for each cluster revealed that Cluster 1 generated 10 significant association rules with a minimum support of 0.01 and confidence of 0.7, while Clusters 2, 3, and 4 produced 2, 3, and 4 association rules, respectively, with adjusted parameters. All rules showed a lift value > 1, indicating positive relationships between products. The findings of this study provide strategic insights for companies in designing data-driven marketing approaches, such as more targeted product offerings for loyal customers or retention strategies for passive customers, thereby optimizing sales and increasing profitability in each customer segment.