Publication Search

The study presents a comprehensive framework for optimizing customer retention budget by integrating clustering, classification, and mathematical optimization techniques. The study begins with the IBM Telco dataset, which is prepared through data cleansing, encoding, and scaling.  In the preliminary phase, customer segmentation is performed using K-Means clustering, with k = 3 and k = 4 identified as optimal based on the elbow method and Silhouette score. The configurations produced three (Premium, Standard, Low) and four (Premium, Standard Plus, Standard, Low) customer segments based on purchase preferences, which served as input features for churn prediction. In the second phase, the dataset was divided into training and test sets in an 80:20 ratio, followed by data balancing using the Synthetic Minority Over-sampling Technique (SMOTE) and Edited Nearest Neighbors (ENN). Multiple classification algorithms were evaluated, including Naive Bayes (NB), Random Forest (RF), Categorical Boosting (CatBoost), Light Gradient Boosting Machine (LightGBM), Extreme Gradient Boosting (XGBoost), Gradient Boosting (GB), Support Vector Machine (SVM), Logistic Regression (LR), K-Nearest Neighbors (KNN), and Multi-Layer Perceptron (MLP) using F1-score as the performance metric. CatBoost and LightGBM, with k values of 3 and 4, respectively, were the highest-performing classification models, with only minimal differences in performance.    Ultimately, customer segmentation established customer prioritization, whereas churn prediction assessed customer churn likelihood. Four distinct configurations were assessed utilizing mixed-integer linear programming (MILP) to optimise retention budget allocation within uniform budget constraints, discount amounts, and churn thresholds. In both the k=3 and k=4 scenarios, CatBoost surpassed LightGBM, with CatBoost at K=3 effectively discounting 66% of at-risk consumers across all three segments, hence improving the intervention's efficacy and budget allocation, making it the ideal choice for maximizing customer retention. The results demonstrate the importance of segmentation in enhancing retention budgeting and budget optimization, particularly concerning parameter sensitivity.

SPP Bandung plays a role in managing the distribution of shipments effectively and efficiently. SPP Bandung handles two types of distribution, namely primary distribution outside the Bandung area and secondary distribution to 12 distribution centers within the Bandung area. Currently, SPP operates six fleets to serve both types of distribution. However, the current distribution system has not optimized the fleet capacity, resulting in a waste of time and resources. This study uses a quantitative approach to analyze and develop an optimal distribution route strategy. By applying the saving matrix, nearest insertion, and nearest neighbor methods, the analysis results showed a decrease in total mileage from 261.8 km to 249.6 km and fuel cost efficiency from Rp210,377.33 to Rp198,539.33. The aving matrix method proved to be effective in route optimization, but other methods can also be considered for maximum improvement.

Gangguan tidur pada mahasiswa merupakan permasalahan yang dapat berdampak pada kesehatan jantung, khususnya melalui perubahan aktivitas listrik jantung yang terekam dalam sinyal EKG. Penelitian ini bertujuan mengembangkan sistem klasifikasi otomatis untuk mendeteksi kondisi jantung berdasarkan sinyal EKG menggunakan algoritma K-Nearest Neighbor (KNN) dan reduksi fitur dengan Principal Component Analysis (PCA). Dataset yang digunakan terdiri dari 159 citra sinyal EKG yang dibagi menjadi dua kelas, yaitu Good Heart dan Bad Heart. Citra diproses melalui tahap preprocessing, reduksi dimensi menggunakan PCA, dan diklasifikasikan menggunakan KNN dengan berbagai nilai parameter. Model terbaik diperoleh pada kombinasi 20 komponen PCA dan nilai K = 6, dengan akurasi mencapai 96,23%, precision 98,46%, recall 92,75%, dan f1-score 95,50%. Hasil penelitian menunjukkan bahwa metode ini mampu mengenali kondisi jantung dengan baik dan efisien. Sistem ini berpotensi dikembangkan sebagai alat bantu deteksi dini gangguan jantung, khususnya pada kelompok mahasiswa yang mengalami gangguan tidur.

Tingkat keberhasilan deteksi penyakit jantung sangat bergantung pada akurasi model klasifikasi yang digunakan. Penelitian ini bertujuan membandingkan kinerja dua algoritma klasifikasi, yaitu K-Nearest Neighbor (KNN) dan Support Vector Machine (SVM), dalam mendeteksi penyakit jantung menggunakan dataset berjumlah 1025 sampel dengan dua kelas target, yakni sehat dan penyakit jantung. Proses pra-pemrosesan data meliputi pembersihan dan normalisasi fitur medis seperti usia, tekanan darah, serta kadar kolesterol. Evaluasi performa model dilakukan menggunakan metode Confusion Matrix, K-Fold Cross Validation, kurva Receiver Operating Characteristic (ROC), dan kurva Precision-Recall untuk mengukur akurasi, presisi, recall, serta keseimbangan antara presisi dan recall. Hasil pengujian menunjukkan bahwa algoritma KNN unggul dalam menghasilkan akurasi tinggi yaitu 99% dengan AUC ROC sempurna 1.00 dan presisi yang hampir konsisten sepanjang recall, sementara SVM menunjukkan performa stabil dengan akurasi 91%, AUC ROC 0.97, dan AP Precision-Recall sebesar 0.96. Penelitian ini menegaskan efektivitas KNN dalam menghasilkan prediksi penyakit jantung yang sangat akurat dengan potensi risiko overfitting pada parameter k kecil, sedangkan SVM memberikan kestabilan model dengan kemampuan generalisasi yang lebih baik. Temuan ini diharapkan dapat menjadi referensi dalam pemilihan algoritma klasifikasi yang sesuai untuk mendukung diagnosis penyakit jantung secara klinis.

This study aims to conduct customer segmentation and develop a classification model to predict the clusters of new customers at Monex Toys Abadi Bekasi, a micro, small, and medium enterprise (MSME). Segmentation was performed using the K-Means Clustering algorithm, incorporating parameters such as Recency, Frequency, Monetary (RFM), purchased products, payment methods, shipping cost discounts, and the total number of products purchased by customers. The segmentation results revealed two clusters: (1) Discount Hunters and (2) Loyal Customers. Subsequently, a classification process was conducted to predict customer clusters using the K-Nearest Neighbor (KNN) and Support Vector Machine (SVM) algorithms. Evaluation results indicated that all models achieved high accuracy exceeding 98%. The best-performing model was obtained with SVM using a 70:30 data split, achieving an accuracy of 98.81%. This classification model was then implemented into a Streamlit-based cluster prediction application, enabling users to identify customer segments in real-time. The findings of this research are expected to assist MSMEs in understanding customer behavior, enhancing service quality, and supporting more effective marketing strategies.

The utilization of Artificial Intelligence (AI) in healthcare services and medical diagnosis in Indonesia has grown rapidly alongside the digital transformation of the health sector. AI technology has been employed to improve service efficiency, accelerate diagnostic processes, and enhance disease detection accuracy, particularly through medical imaging and ECG data analysis. Algorithms such as K-Nearest Neighbor (KNN) and Chi-Square have shown effectiveness in heart disease classification. However, despite its benefits, AI implementation presents legal challenges. The absence of specific regulations regarding legal liability in cases of AI-based diagnostic errors creates uncertainty for both medical professionals and patients. Additionally, the lack of national standards, weak patient data protection, and digital literacy gaps present significant obstacles. Adaptive policies, the establishment of dedicated regulations, and collaboration between government, medical practitioners, technology developers, and academics are essential to develop a legal framework that accommodates AI advancements responsibly. With clear legal certainty, AI technology can be optimally utilized to support more inclusive and high-quality healthcare services.

This study aims to analyze and implement the K-Nearest Neighbor (KNN) algorithm in determining the feasibility of sheep selection at PT Arjuna Farm. In the livestock industry, selecting quality animals is very important to increase productivity and efficiency. KNN, as a simple but effective classification method, is used to analyze sheep characteristic data, such as body weight, age, and health, to identify animals that meet the eligibility criteria. The research methodology used in this study includes data collection, pre-processing, and application of the KNN algorithm. Data obtained from PT Arjuna Farm were processed and divided into training data and testing data. The results of the analysis show that the KNN algorithm is able to provide high accuracy in determining the feasibility of sheep selection, with a low error rate.

Learners are continually faced with choosing appropriate courses or making career choices due to increased educational opportunities. The emergence of machine learning-based course and career recommender systems has the potential to address this issue, offering personalized course recommendations tailored to individual learning pathways, preferences, and learning history. The optimization and feature engineering techniques and practical deployment environments have not been collectively examined in the previous research, despite the significant advancements in this area of research. Furthermore, previous research has rarely synthesized how these technical components help students choose appropriate courses and careers. This systematic review was carried out to investigate the current state of machine learning-based course and career recommender systems, focusing on key elements, such as primary data sources, feature engineering methods, algorithms, optimization techniques, evaluation metrics, and the environments where the existing course recommendation models are deployed. The PRISMA method for conducting a systematic review was used to choose studies that met the requirements for inclusion and exclusion. The study findings show significant reliance on interpretable and traditional machine learning algorithms, such as K-Nearest Neighbor and Random Forest, to develop recommender models. Feature engineering remains basic, as most studies rely on normalization, while optimization processes are often underreported. Also, evaluation metrics varied widely, impeding comparability, while most of the recommender models are deployed in an e-learning environment, leaving the traditional learning environment underrepresented. Furthermore, the study findings identified issues including data sparsity and diversity, data security and privacy, and changes in learner preferences that may have an impact on the performance of recommender systems while recommending further studies to make use of standardized optimization methods, and automated domain-informed feature engineering frameworks, benchmark and annotated datasets in developing models the gives priority to learners’ success and educational relevance.

Coronary heart disease (CHD) remains a leading cause of mortality worldwide. Early detection is essential to reduce complications and improve patient outcomes. This study aims to develop a classification model using machine learning algorithms to predict CHD risk based on clinical symptoms. The dataset used is the Cleveland Heart Disease dataset from the UCI Machine Learning Repository, consisting of 303 patient records with 14 clinical features. The preprocessing stage involved handling missing values, normalizing features, and transforming categorical variables. Four classification algorithms were applied: K-Nearest Neighbors (K-NN), Decision Tree, Random Forest, and Support Vector Machine (SVM). Each model was trained using stratified 10-fold cross-validation to ensure generalizability. Evaluation using accuracy, precision, recall, F1-score, and ROC-AUC metrics showed that the Random Forest algorithm achieved the highest performance with 87.2% accuracy. Feature importance analysis indicated that chest pain type, resting blood pressure, cholesterol, and ST depression were the most influential indicators. These results demonstrate that machine learning, particularly Random Forest, can effectively support early diagnosis of CHD in clinical settings and has the potential to be integrated into clinical decision support systems (CDSS).

Given the increasing prevalence of mental health challenges in digital work settings, especially among IT remote workers, early detection mechanisms have become critically important. This study aims to improve the prediction accuracy of mental health conditions among IT remote workers by integrating feature engineering techniques within machine learning models. Five algorithms consisting of Random Forest, Logistic Regression, K-Nearest Neighbors, Decision Tree, and Naive Bayes were evaluated. The Random Forest model achieved the best performance, with 83% accuracy, 83% precision, 100% recall, and a 90% F1-score, followed closely by Logistic Regression with 82% accuracy. Nevertheless, the results demonstrate the feasibility of applying machine learning to support the early detection of mental health risks, offering a strong foundation for future research in predictive analytics and the development of intelligent support systems within digital work environments.

The online gaming industry continues to grow rapidly in Indonesia, with many users purchasing digital items through 3rd party top up services such as Pitopup.com. One of the main challenges faced by Pitopup.com is the difficulty in classifying the sales of each available game item. This research aims to apply the K-Nearest Neighbor (KNN) method to predict the sales classification of game items in order to find out the sales category for each game item and hopefully help increase stock efficiency. The dataset used was obtained from historical sales data on Pitopup.com from June to September 2024. The research stages include data processing, normalization using Min-Max Scaling, data transformation using label encoding, separating test and training data using a ratio of 80:20, and using confusion matrix as a model evaluation. The test results show that KNN algorithm is able to classify game item sales on the Pitopup.com website with a level of accuracy in several categories: marketable category at 100%, the moderately sellable category at 100% and the not sellable category at 100%.

Optimizing distribution routes is an important step for MSMEs in increasing operational efficiency and customer satisfaction. This research was conducted on Bolen Crispy Mak Tin MSMEs which face distribution challenges with routes that are not yet optimal, causing increased transportation costs and the risk of decreasing product quality. This research uses the Branch and Bound and Nearest Neighbor algorithms to solve the Traveling Salesman Problem (TSP) problem in determining efficient distribution routes. The results of data processing are optimal routes that have the minimum distance with a total distance of 282.5 KM with route P-1-2-7-4-5-6-3-0 for the branch and Bound algorithm and 239 km with route P- 2-3-4-5-6-7-1-P for Nearest Neighbor This result is more optimal when compared to the previous route, namely P-1-2-3-4-5-6-7-P with a distance of 291 km analysis shows that Method Nearest Neighbor is able to provide an optimal solution by minimizing travel distance and distribution costs, while the Branch and Bound algorithm also provides an optimal solution but is less efficient. and distribution cost efficiency from Rp. 570,320.9 to 565,511.36 or 0.84% ​​more savings for the Branch and Bound Algorithm and 540,795.45 or 5.18% more savings for Nearest Neighbor

Due to the variety of types of flowers that exist and having and tracking each variety, making plant lovers and cultivators difficult to distinguish in determining the type of flower, it takes a very long time to find out the type of flower if you only rely on the five senses. With the application of the K-Nearest Neighbor algorithm and feature extraction of color and texture, it is very helpful in image processing to identify flowers more easily and shorten the time, with the greatest accuracy of 71% using the K-7 value, the flower was successfully carried out.

Advances in information and internet technologies have significantly transformed the business environment, including the financial sector. The COVID-19 pandemic has further accelerated this digital adoption, expanding the e-commerce industry and highlighting the necessity for secure online transactions. Credit Card Fraud Detection (CCFD) stands critical as the prevalence of fraudulent activities continues to rise with the increasing volume of online transactions. Traditional methods for detecting fraud, such as rule-based systems and basic machine learning models, tend to fail to keep pace with fraudsters' evolving tactics. This study proposes a novel ensemble deep learning-based approach that combines Convolutional Neural Networks (CNN), Gated Recurrent Units (GRU), and Multilayer Perceptron (MLP) with the Synthetic Minority Oversampling Technique and Edited Nearest Neighbors (SMOTE-ENN) to address class imbalance and improve detection accuracy. The methodology integrates CNN for feature extraction, GRU for sequential transaction analysis, and Multilayer Perceptron (MLP) as a meta-learner in a stacking framework. By leveraging SMOTE-ENN, the proposed approach enhances data balance and prevents overfitting. With synthetic data, the robustness and accuracy of the model have been improved, particularly in scenarios where fraudulent examples are scarce. The experiments conducted on real-world credit card transaction datasets have established that our approach outperforms existing methods, achieving higher metrics performance.

This study aims to optimize the implementation of the K-Nearest Neighbors (K-NN) algorithm for medical image classification by focusing on selecting the optimal KKK parameter and applying dimensionality reduction techniques to improve accuracy and efficiency. The data used was sourced from public medical image repositories such as The Cancer Imaging Archive (TCIA) and Medical Image Analysis datasets, covering various diseases, including brain tumors, lung cancer, and kidney lesions. The research process involves data collection, data preprocessing, dimensionality reduction using Principal Component Analysis (PCA), applying the K-NN algorithm with Euclidean, Minkowski, and Cosine distance metrics, and performance evaluation using accuracy, precision, recall, and F1-score. Experimental results demonstrate that K=5with the Euclidean distance metric provides the best performance, achieving an accuracy of 90%. Additionally, PCA effectively reduces computational time by 30% without significantly compromising accuracy. This study proves that K-NN is an effective method for medical image classification. However, further research is needed to integrate K-NN with deep learning models to enhance performance and feature extraction capabilities.

This study conducts an exploratory data analysis combined with machine learning techniques to identify early signs of student depression. We investigated various factors affecting mental health among students, including sleep duration, dietary patterns, history of suicidal thoughts, family history of mental illness, and their relationships with depression across age groups and academic pressure. The study also examined the influence of gender on academic stress levels. Three machine learning models such as Random Forest, Support Vector Machine (SVM), and K-Nearest Neighbors (KNN) were utilized to predict depression. The performance of these models was evaluated, achieving accuracy rates of 84.97% for Random Forest, 84.85% for SVM, and 81.16% for KNN. The findings highlight the effectiveness of these models in predicting student depression and underscore the importance of targeted mental health interventions based on key factors influencing mental health among students.

Abstract. From the perspective of Micro, Small, and Medium Enterprises (MSMEs), fluctuations in raw material prices are highly concerning as they can significantly impact business stability. While MSMEs may tolerate price fluctuations to some extent, from an industrial engineering perspective, such a passive approach contradicts the principles of continuous improvement. This study seeks to predict the price of large red chili peppers using five regression models implemented through Orange Data Mining: Linear Regression, Support Vector Machine, Decision Tree, k-Nearest Neighbors (kNN), and Gradient Boosting. Due to the limited availability of daily data, particularly within a daily timeframe, the study utilized weekly data spanning three years. The results of the Test and Score evaluation shows Gradient Boosting as the best-performing model, achieving a Mean Absolute Percentage Error (MAPE) of 0.7%. However, the MAPE for predictions in January 2025 increased to 15.8%. This error is expected to decrease as more weekly data becomes available to mitigate the inaccuracies inherent in this model. Keywords: prediction, red chilli, regression, supervised learning , orange data mining. Abstrak. Dalam perspektif UMKM, fluktuasi harga bahan baku adalah suatu hal yang paling ditakuti karena berakibat pada ketahanan usaha yang menjadi tidak menentu. Pada suatu kondisi, fluktuasi harga dapat diterima para UMKM, namun dalam perspektif teknik industri, sikap UMKM tersebut tidak sesuai prinsip continuous improvement. Penelitian ini mencoba untuk memprediksi harga cabai merah besar dengan menggunakan 5 model regresi dibantu Orange Data Mining. Yaitu Linear Regression, Support Vector Machine, Tree, kNN, Gradient Boosting. Data yang diperlukan sebagian besar tidak tersedia, khususnya dalam kerangka waktu harian sehingga penelitian ini menggunakan data mingguan selama 3 tahun. Hasil Test and Score menunjukkan model Gradient Boost terpilih menjadi model terbaik dengan tingkat MAPE 0.7% namun MAPE pada tahap Prediction di bulan Januari 2025 menjadi 15.8%. Error tersebut akan berkurang ketika data mingguan sudah cukup banyak untuk menambal kesalahan yang dihasilkan model ini Kata kunci: prediksi, cabai merah, regression, supervised learning , orange data mining.

In today's era, there are already many companies that have been established, from urban to rural areas, various companies have been established, especially companies that provide daily necessities such as supermarkets. And each company competes with each other in selling its products with the expected results. In this study, researchers use data to support this study. Because sales of goods or product stock can be calculated in sales results, the higher the sales, the higher the risk that will be faced. This study aims to apply data mining in analyzing sales results that occur in supermarkets. And to find out its impact on sales. This researcher uses the KNN method, by looking for test results with this method which will be implemented using the Rapid Miner application which will later produce the results of its analysis.

In today's era, there are already many companies that have been established, from urban to rural areas, various companies have been established, especially companies that provide daily necessities such as supermarkets. And each company competes with each other in selling its products with the expected results. In this study, researchers use data to support this study. Because sales of goods or product stock can be calculated in sales results, the higher the sales, the higher the risk that will be faced. This study aims to apply data mining in analyzing sales results that occur in supermarkets. And to find out its impact on sales. This researcher uses the KNN method, by looking for test results with this method which will be implemented using the Rapid Miner application which will later produce the results of its analysis.

In today's world of retail and technology, competition is fiercely competitive. With the development of retail businesses increasing in number and mushrooming in a region, consumer needs are increasing, and retail business players are competing to develop their businesses by utilizing existing technology. Daily sales transaction data continues to increase, causing a lot of storage. Toko Ira has more than 228 sales transaction data records from 2023 to 2024 that have not been used. Data requires a lot of storage space. Additionally, the data has not been used in an effective way. Based on this problem, this research aims to use data mining to classify sales transaction data to determine which items are selling best. This research is a case study with a qualitative approach. This research was conducted with the Naive Bayes method and Rapidminer was used. The results of the sales transaction data classification research are the division of products into best-selling and non-selling categories. The results of this research show that the K-Nearest Neighbors (KNN) algorithm with a 50:50 data division is more effective in predicting and classifying sales of best-selling and non-selling products in IRA stores. The results show that the Naive Bayes algorithm has an accuracy of 89.91%, while the K-Nearest Neighbors (KNN) algorithm has an accuracy of 60.09%.