MEANS (Media Informasi Analisa dan Sistem)

Penerapan Algoritma K-Means Dalam Klasterisasi Penjualan Obat Pada Apotek Kharisma Farma

2025-03-17T05:54:15+01:00

Kharisma Farma Pharmacy sells various over-the-counter medicines, herbals, patents, external medicines, syrups, generics and other health devices. The average drug sales transaction in 2024 is 150-250 transactions per month. The results of interviews with the pharmacy, drug sales data are only used for calculating gross, net and archive profits without ever being analyzed, even though the data can identify sales trends that can be used to ensure product availability according to demand so that the pharmacy can optimize stock and reduce the risk of shortages and excess stock of drugs. The analysis process uses the K-Means algorithm. The results of applying the K-Means Algorithm can determine the trend of drug sales in June and July 2024 into 5 clusters, cluster 1 (very popular) with 16 drug items, cluster 2 (popular) with 5 drug items (quite popular), cluster 3 with 5 drug items, cluster 4 (less popular) with 9 drug items, cluster 5 (very less popular) with 16 drug items.

Implementasi Algoritma K-Means untuk Clustering Nominal Pembayaran pada Odoo Versi 11 di PT.XYZ

2025-11-17T02:58:48+01:00

Payment data in the Odoo version 11 system at PT. XYZ has not been optimally utilized to support strategic business decisions. This study implements the K-Means clustering algorithm to group customer payment amounts and identify transaction behavior patterns. The optimal number of clusters was determined using the Elbow Method and Silhouette Score, while visualization through scatter plot and block plot was used to interpret the clustering results. The analysis produced three main clusters representing customers with high, medium, and low payment amounts. These segmentation results enable PT. XYZ to better understand customer payment behavior, optimize marketing strategies, enhance service quality, and improve cash flow management. Overall, this research demonstrates the effectiveness of the K-Means algorithm in processing payment data within the Odoo ERP system and highlights its potential to support more accurate and data-driven business decision-making.

Implementasi K-means untuk Clustering pada Industri Non-Agro di Jawa Tengah

2026-01-02T02:06:39+01:00

This study aims to segment the non-agro industry in Central Java Province using the K-Means Clustering method. The approach used refers to the Knowledge Discovery in Database (KDD) stages, which include data selection, preprocessing, transformation, data mining, and interpretation/evaluation. Data were obtained from the Central Java Open Data portal, with variables used including district/city, industry type, business scale, marketing reach, and energy use. The preprocessing stage was carried out through data cleaning, integration of several regional files, and normalization using StandardScaler. The data mining process was carried out using the K-Means algorithm and the determination of the optimal number of clusters using the Elbow Method. The results of the study indicate the formation of several clusters that describe market segmentation patterns in the non-agro industry based on similarities in operational characteristics between regions in Central Java.

Analisis Sistem Antrian Kendaraan Bermotor Menggunakan Metode Discrete-Event Simulation pada SPBU 14.2031143 Tanjung Anom Deli Serdang

2025-12-22T06:26:52+01:00

This study aims to analyze the performance of the motor vehicle queuing system at gas station 14.2031143 Tanjung Anom Deli Serdang using the Discrete-Event Simulation (DES) approach. Data was obtained through direct observation for 1 hour and 10 minutes by recording the arrival time, start of service time, and end of service time for each vehicle. The results of data processing show that the average waiting time for vehicles reached 5 minutes and 42 seconds, which is much higher than the average service time of 25 seconds. This condition indicates that the main problem lies in the density of the queue lane, not in the service process. Based on these results, a DES simulation model was developed to represent the actual behavior of the queueing system and used as the basis for evaluating improvement scenarios. The main scenario tested was the addition of parallel queuing lanes to reduce vehicle congestion in a single line. The simulation results showed that the addition of lanes significantly reduced waiting times and improved the smooth flow of the queue without requiring changes to service facilities. This study concludes that adding queue lanes is an effective and feasible solution to improve the efficiency of the queue system at gas stations, and can be implemented as a measure to improve service during busy operating hours.

Perbandingan Kinerja Algoritma Apriori dan Fp-Growth dalam Analisis Pola Pembelian Produk pada Timmy Store

2025-11-21T05:30:38+01:00

Pemanfaatan teknik data mining dalam industri ritel digital semakin dibutuhkan untuk memahami pola perilaku konsumen. Salah satu pendekatan yang banyak digunakan adalah Market Basket Analysis (MBA), yang bertujuan mengenali keterkaitan antarproduk dalam transaksi. Penelitian ini membandingkan kinerja dua algoritma MBA, yaitu Apriori dan FP-Growth, dalam menganalisis pola pembelian produk di Timmy Store menggunakan data transaksi periode Desember 2024-Maret 2025. Setelah dilakukan proses pembersihan, diperoleh 2.274 transaksi valid yang kemudian dibentuk menjadi 220 keranjang pembelian berdasarkan identitas pelanggan. Hasil penelitian menunjukkan bahwa algoritma Apriori mampu menghasilkan 15 frequent 1-itemset dan 4 frequent 2-itemset, serta tiga aturan asosiasi yang memenuhi batas minimum support 5% dan confidence 50%. Aturan paling kuat ditemukan pada keterkaitan Ancient Megalodon Fisch Phantom Megalodon dengan support 0,159, confidence 62,5%, dan lift 1,88. Sementara itu, algoritma FP-Growth menghasilkan pola yang sama, tetapi dengan waktu proses yang lebih cepat akibat tidak dilakukannya pembangkitan kandidat seperti pada Apriori. Secara keseluruhan, FP-Growth lebih unggul dalam aspek efisiensi komputasi, sedangkan Apriori tetap relevan karena memberikan transparansi perhitungan yang lebih mudah ditelusuri.

Rancang Bangun Sistem Informasi Sebagai Media Promosi Dengan Metode RAD Pada UMKM Kopi Bintang Berbasis Website

2025-10-02T04:33:39+02:00

technology has now penetrated many sector, including the culinary industry. One particularly effective strategy is the use of online technologi through websites. Content displayed on digital pages can make food and beverage products more attractive, while simultaneously opening up business opportunities to reach more customers. Website also provide a more communicative and interactive promotional space. Kopi bintang is one example of a culinary business focused on processed food and beverage products. In the initial stages, the promotional strategy used relied solely on social media, but these efforts did not yield optimal results in attracting consumers, there fore, innovation was needed in the form of a website that no tonly supports promotions but also supports business processes more effectively. This research was conducted using interviews, field observations, and desk research. After data collection, the system was designed using the rapit application development (RAD) approach. System evaluation was conducted through black bos testing to ensure each feature operated as intended and met its objectives. The resulting output was a website design that server a promotional medium for the kopi bintang MSME.

Sistem Chatbot Helpdesk WhatsApp Berbasis Large Language Model Gemini dengan Fine-Tuning Menggunakan Retrieval-Augmented Generation dan Qdrant Vector Database (Studi Kasus: Universitas Multi Data Palembang)

2026-01-08T02:58:25+01:00

Universitas Multi Data Palembang encounters significant challenges in handling a high volume of student inquiries via WhatsApp, resulting in prolonged response times and limited service availability outside operational hours. This study aims to design and implement an intelligent 24-hour helpdesk chatbot system to enhance academic service efficiency. The research employs the Research and Development (R&D) method. The system architecture integrates the Google Gemini Large Language Model (LLM), with performance optimized through the Retrieval-Augmented Generation (RAG) approach and Qdrant Vector Database. This technique enables the chatbot to access an internal knowledge base constructed from official university documents in real-time, thereby minimizing hallucinations. Performance evaluation using the RAGAS framework demonstrates a substantial improvement in Answer Correctness, achieving a score of 0.89, and a Faithfulness score of 0.98 compared to the model without RAG. Furthermore, User Acceptance Testing (UAT) involving 64 respondents yielded an average score of 4.26 (Very Good Category), indicating that the system is feasible for implementation to support rapid and accurate academic information services.