Figure 1 - available via license: Creative Commons Attribution-NonCommercial 4.0 International
Content may be subject to copyright.
Source publication
RFM stands for Recency, Frequency, and Monetary. RFM is a simple but effective method that can be applied to market segmentation. RFM analysis is used to analyze customer’s behavior which consists of how recently the customers have purchased (recency), how often customer’s purchases (frequency), and how much money customers spend (monetary). In thi...
Similar publications
This study validated the CO2 distribution predicted by a computational fluid dynamics model considering CO2 absorption by photosynthesis in a chamber and greenhouse. The effect of photosynthesis with CO2 emission from a perforated tube remains not fully understood, although previous studies on CO2 distribution in greenhouses have been conducted. Mo...
Increasing users' positive interactions, such as purchases or clicks, is an important objective of recommender systems. Recommenders typically aim to select items that users will interact with. If the recommended items are purchased, an increase in sales is expected. However, the items could have been purchased even without recommendation. Thus, we...
We compare the impact of two different listing price strategies for residential homes on the purchasing price of a property. Previous literature on anchoring effect (Tversky and Kahneman in Science 185(4157):1124–1131, 1974) has encountered a direct relation between the listing price and the sale price. Among the listing prices, the asking price, p...
The rising number of micro-businesses has, in one way or another, created an impact on the Philippine economy. However, the sustainability of these micro-businesses seems to be in question, for they do not last long in the industry. Hence, the entrepreneurial orientation of micro-businesses was investigated, and its impact on the business's sales g...
We consider the use of P-spline generalized additive hedonic models for real estate prices in large U.S. cities, contrasting their predictive efficiency against linear and polynomial based generalized linear models. Using intrinsic and extrinsic factors available from Redfin, we show that GAM models are capable of describing 84% to 92% of the varia...
Citations
... Simply dropping any of these features would result in the loss of relevant information. Principal Component Analysis (PCA) is utilized to reduce the dimensionality of the variables [18]. It was decided to evaluate using two of these principal components, which together account for 85.3% of the explained variance. ...
... Authors in [10] emphasized that the RFM model can also induce crucial insights for market clustering. The RFM model can be used to develop effective marketing strategies [11] and can be effectively implemented in the sales sector [12,13]. ...
Market clustering is increasingly important for companies to understand consumer shopping behavior in the context of complex data. This study aims to develop a hybrid model that integrates Principal Component Analysis (PCA) and k-medoids to enhance market clustering based on consumer shopping patterns. The methods used include data preprocessing, PCA application for dimensionality reduction, and clustering using k-medoids. The quality of the clusters is evaluated with various validity indices. The results show that the hybrid model produces clusters with better quality compared to the single k-medoids method, as seen from the Calinski-Harabasz Index (CHI), the Silhouette Width (SW), and the Davies-Bouldin (DB) index. The implications of these findings emphasize the importance of adopting hybrid methods in marketing strategies to improve understanding of consumer behavior dynamics and allow companies to adjust their marketing strategies more effectively. This study provides a strong foundation for further development in clustering analysis across various industry sectors and highlights the potential for innovative techniques to address dynamic market challenges.
... RFM analysis methods can be classified into three types. Methods of the first type divide customers by clustering, such as using the -means algorithm [25], which is the most commonly used approach. It aims to determine customer value and loyalty by clustering customers based on their characteristics. ...
In recent years, data mining technologies have been well applied to many domains, including e-commerce. In customer relationship management (CRM), the RFM analysis model is one of the most effective approaches to increase the profits of major enterprises. However, with the rapid development of e-commerce, the diversity and abundance of e-commerce data pose a challenge to mining efficiency. Moreover, in actual market transactions, the chronological order of transactions reflects customer behavior and preferences. To address these challenges, we develop an effective algorithm called SeqRFM, which combines sequential pattern mining with RFM models. SeqRFM considers each customer's recency (R), frequency (F), and monetary (M) scores to represent the significance of the customer and identifies sequences with high recency, high frequency, and high monetary value. A series of experiments demonstrate the superiority and effectiveness of the SeqRFM algorithm compared to the most advanced RFM algorithms based on sequential pattern mining. The source code and datasets are available at GitHub https://github.com/DSI-Lab1/SeqRFM.
... Using k-means clustering algorithm, optimal silhouette core was achieved with k = 2. This value indicates that air pollutant pro le in each target location can be distinctively clustered in two groups (Gustriansyah et al., 2020;Zhang et al., 2021). Figure 2 shows the cluster box plot comparison for the most relevant pollutants which were found to be NO 2 , PM 2.5 , PM 10 , and NMHC. ...
Air pollution poses significant risks, particularly in developing countries where rapid urbanization exacerbates pollutant emissions. These pollutants impact local populations and contribute to global air quality challenges through long-range transport. Despite numerous studies, comprehensive data on pollutant profiles in urban areas remain limited by the availability of stationary monitoring systems. This study addresses these gaps by employing a novel approach that combines mobile in-situ air quality measurements with clustering and back trajectory analysis to map and trace pollution sources across diverse urban environments. The use of mobile instruments allows for resource-efficient data collection, enhancing the ability to identify pollution hotspots without requiring extensive infrastructure. The analysis revealed two distinct pollutant clusters: aerosol pollutants dominated in residential areas, while gaseous pollutants were more prevalent near traffic-heavy and construction areas. This methodology not only provides a scalable alternative to traditional air quality monitoring but also offers valuable insights into pollution source attribution, particularly in developing countries where resources for environmental monitoring are limited.
... (48) states that a user is provided either the wireless or the fiber connection. (49) states that a fiber user is connected to either a PPS or a SPS. (50) states that if a user is provided the wireless connection, then it is implemented using one of the beams of its associated sector of the associated gNB. ...
... As observed from Figure 7, the absence of an always-increasing trend in CH value signifies that increasing the number of clusters will not always lead to better results. Since the CH index considers both intracluster compactness and inter-cluster separation, it is possible that the clustering algorithm may not exhibit a straightforward relationship between CH values and the number of clusters [49] Figure 8 from clustering solution) along with the optical devices (OLT, PPS, and SPS) in the deployment region. Figure 8(b) illustrates the deployment of 29 gNBs in the deployment area and the distribution of 500 users under the coverage of the gNBs. ...
This paper investigates the deployment of a cost-efficient fiber-wireless (FiWi) access network with joint utilization of fixed wireless access (FWA) and fiber-to-the-home (FTTH) technologies. In this work, we consider that residential users over a geographical area are to be provided with network services, constrained by the available network resources. We explore the implementation of hybrid FiWi access network that integrates a fiber-based passive optical network (PON) and fifth-generation (5G) wireless access network to provide efficient network services. This paper proposes a methodology for topology optimization of FiWi access networks, considering the usage of three-dimensional (3D) beamforming and 3D resource grid for downlink transmission from the gNB to the users in 5G scenario. We derive the beam codebook and generate multiple beams to simultaneously serve the spatially separated users. We derive the closed-form expression for millimeter-wave channel model incorporating large-scale and small-scale parameters to compute the effective SINR of users. Further, we propose an optimization framework for optimal resource allocation (i.e., beam and resource block) by utilizing the 3D resource grid for downlink transmission. We perform extensive simulations to demonstrate the effectiveness of the proposed methodology for various 3GPP 5G outdoor propagation scenarios, viz., RMa, UMa, and UMi-street canyon.
... Simply dropping any of these features would result in the loss of relevant information. Principal Component Analysis (PCA) is utilized to reduce the dimensionality of the variables [18]. It was decided to evaluate using two of these principal components, which together account for 85.3% of the explained variance. ...
Having a ready-to-use dependent churn variable in the non-subscription based e-commerce domain is not straightforward, as customers do not formally cancel their status as a customer but simply stop placing orders. In this study, a k-means clustering algorithm with RFM variable segmentation is used to assist in classifying customers as churned or not churned. Next, a novel approach is adopted by transforming tabular customer data into radar chart images which are then fed into a pre-trained Vision Transformer model for supervised image classification. The results of the study indicate that while the Vision Transformer model show significant increase in precision and F1-score evaluation met-rics, its performance in MCC, recall, and AUCROC is comparable to that of Cat-Boost and XGBoost. This study demonstrates the potential of transforming tab-ular data into images for training Vision Transformers, for use in customer churn prediction within the non-subscription-based e-commerce domain.
... CHI mengukur rasio pemisahan berdasarkan jarak maksimum antara centroid, dan mengukur kekompakan berdasarkan jumlah jarak antara setiap data dengan centroid. Konfigurasi klaster yang diinginkan adalah yang memiliki varians antar-klaster yang tinggi dan varians intra-klaster yang relatif rendah, menunjukkan bahwa klaster terpisah dengan baik dan kompak secara internal [15]. Rumus untuk mencari Calinski-Harabasz Index, seperti persamaan (5): ...
Museum Monpera Palembang adalah sebuah museum yang memiliki koleksi foto-foto pahlawan nasional Indonesia. Koleksi foto tersebut memiliki nilai historis dan makna yang mendalam bagi masyarakat Indonesia, tetapi beberapa di antaranya sudah terlihat samar dan kabur sehingga informasi yang tergambar menjadi tidak jelas. Penelitian ini bertujuan untuk menerapkan metode klasterisasi pixel yang digabungkan dengan metode K-Means pada aplikasi Augmented Reality untuk melakukan pengujian kualitas citra pada koleksi foto pahlawan museum Monpera. Penelitian ini menggunakan algoritma K-Means yang merupakan salah satu algoritma partitional yang didasarkan pada penentuan jumlah awal kelompok dengan mendefinisikan nilai centroid awalnya. Penelitian ini juga menggunakan teknologi Augmented Reality berbasis Android untuk memberikan pengalaman interaktif kepada pengunjung museum. Hasil pengujian citra menggunakan metode K-Means menunjukkan data evaluasi yang melibatkan Silhouette Score, Calinski-Harabasz, dan Dunn Index. Hasil pengujian ini menunjukkan bahwa metode K-Means belum mampu meningkatkan kualitas citra hasil klasterisasi pixel, tetapi penelitian ini berhasil mengembangkan aplikasi AR dan memberikan kontribusi penting dalam memahami dan mengatasi tantangan dalam mempertahankan integritas visual dari koleksi foto pahlawan nasional Indonesia melalui pengembangan teknik pengolahan citra yang lebih efektif dan inovatif menggunakan metode Clustering pixel dan K-Means dalam konteks Augmented Reality.
... This allows companies to conduct careful procurement of goods, aiming for efficiency. Clustering methods prove highly beneficial in determining the appropriate number of clusters for sales data, aiding in segmenting products effectively based on demand patterns and optimizing inventory management strategies accordingly [5] [6].In this context, clustering analysis emerges as a useful tool for grouping inventory items. Achieving optimal inventory levels presents a significant http://ijstm.inarah.co.id 674 challenge, as excessive inventory can lead to high costs, while inventory that is too low can result in stockouts, potentially leading to lost sales and dissatisfied customers [7]. ...
Inventory forecasting is crucial in effective supply chain management and cost reduction. However, traditional forecasting techniques face significant challenges due to the complexity and variability of demand patterns. This study explores the use of K-means clustering, a data-driven approach that can improve inventory forecasting accuracy. By grouping inventory items based on their unique demand profiles, we can create personalized forecasting models for each cluster. This technique enhances demand estimation, helping businesses make informed decisions and optimize their inventory. Our research delves into the use of K-means clustering to identify patterns and similarities within historical demand data. This clustering process divides inventory items into groups with similar demand characteristics. By applying specific forecasting models to each cluster, we achieve greater precision in our predictions. The proposed methodology is rigorously evaluated using real-world inventory datasets, and the results demonstrate its significant superiority in forecasting accuracy compared to conventional non-clustered methods. This study offers compelling evidence and valuable insights for practitioners seeking to improve their inventory management practices through data-driven techniques.
... 6)If the centroid does not change anymore, the clustering process is complete. One of the main problems of the k-Means method is how to determine the optimal number of clusters k [26]. ...
This study explores the application of machine learning for local market prediction in e-commerce. By leveraging the RFM segmentation method, the model predicts product sales based on user shopping patterns. The RFM score, calculated using recency, frequency, and monetary values of customer purchases, segments customers into distinct categories. The research utilizes a dataset obtained through seven parameters and performs data preprocessing. K-Means clustering then classifies customers into Low, Medium, and High levels based on their RFM scores. Customers in the Low category exhibit low purchase activity but high product browsing. The Medium segment displays consistent purchases of a limited product range. High-level customers demonstrate frequent purchases with significant spending. The identified customer segments enable targeted marketing strategies. For Low-level customers, discounts or product feature promotions can incentivize purchases. Combining product offerings can entice Medium-level customers to explore new products. Finally, High-level customers can be engaged through loyalty programs offering rewards. This approach empowers e-commerce sellers to tailor marketing strategies for each customer segment, enhancing market dominance.
... Pengelompokan pada penelitian ini menggunakan salah satu algoritma pada metode clustering yang sesuai untuk mengolah data numerik yaitu Algoritma K-Means. Algoritma K-Means merupakan salah satu metode clustering non-hirarki yang paling membagi data ke dalam cluster tanpa memerlukan label ataupun kategori [6] [7]. Masalah yang dihadapi pada algoritma ini adalah menentukan jumlah cluster yang optimal. ...
... Oleh karena itu, beberapa metode pengujian dilakukan untuk menentukan jumlah cluster di antaranya menggunakan Elbow Method dan Davies Bouldin Index. Kedua metode tersebut digunakan karena dapat mengukur dan memvalidasi dalam menentukan jumlah cluster yang baik dan optimal [6]. ...
... K-Means merupakan metode clustering nonhierarki yang membagi data n ke dalam k cluster, sehingga hasil kesamaan antar clusternya tinggi jika Within-Cluster Sum of Squarenya minimal dan hasil kesamaan di dalam clusternya rendah jika Between-Cluster Sum of Squarenya maksimal [6]. ...
Penelitian ini bertujuan untuk mengelompokkan makanan yang memiliki nilai nutrisi yang serupa. Yang mana makanan dibagi ke dalam 3 cluster yaitu makanan yang mempunyai kadar nutrisi tinggi, sedang dan rendah. Hasil pengelompokan pada penelitian ini dapat digunakan untuk pemilihan dan konsumsi makanan dalam pemenuhan nutrisi dan juga dapat digunakan untuk mencegah timbulnya penyakit yang disebabkan oleh makanan. Seperti makanan pada cluster 0 bisa dipilih jika ingin menaikkan berat badan. Makanan cluster 1 dapat menjadi patokan jika dikonsumsi terlalu banyak dapat menyebabkan obesitas dan cluster 2 dapat dipilih jika ingin melakukan diet atau menurunkan berat badan. Hasil ini ditunjukkan dari hasil klasterisasi di mana cluster pertama diisi oleh makanan dengan kadar kalori dan protein yang cukup tinggi dan kadar lemak, karbohidrat yang rendah. Cluster kedua diisi oleh makanan dengan kadar kalori, protein dan lemak yang tinggi serta kadar karbohidrat yang rendah. Cluster ketiga diisi oleh makanan dengan kadar kalori, protein, lemak dan karbohidrat yang rendah. Penelitian ini menggunakan metode clustering dengan menerapkan algoritma K-Means karena efektif dalam melakukan klasterisasi terhadap tipe data numerik dan menguji menggunakan Elbow Method dan Davies Bouldin Index.
