Abstract:The development of information technology has driven digital transformation in the trade sector, one of which is through e-commerce platforms. In an effort to improve user experience and sales efficiency, this study developed…
loped an e-commerce application that is equipped with a machine learning-based buyer choice analysis feature. The app is designed not only as a buying and selling medium, but also capable of learning user preferences based on interaction data, such as search history, product clicks, and transactions. The machine learning methods used are classification and recommendation algorithms, especially K-Nearest Neighbors (KNN) and Collaborative Filtering, to predict the products that are most likely to be of interest to users. The development process follows the Agile Development model, with stages of UI/UX design, backend system integration, and model accuracy testing. The results of the evaluation showed that the recommended system implemented was able to increase the relevance of the product display by up to 85%, as well as provide a better personalized experience for users. This application is expected to be a smart solution in improving customer satisfaction and the effectiveness of digital marketing strategies
Abstract:Student mental health disorders have become an increasingly important issue due to their significant impact on academic performance and psychological well-being. Machine learning-based approaches offer a promising alternative…
ative for supporting early detection; however, only a limited number of studies have conducted comprehensive evaluations of various ensemble learning algorithms by simultaneously considering predictive performance, computational efficiency, and statistical significance. This study aims to compare ten classification algorithms, namely Random Forest, Extra Trees, Gradient Boosting, Hist Gradient Boosting, AdaBoost, XGBoost, LightGBM, CatBoost, Logistic Regression, and Stacking Ensemble, using 5-fold Stratified Cross-Validation. Model performance was evaluated using Accuracy, Precision, Recall, F1-Score, Receiver Operating Characteristic–Area Under the Curve (ROC-AUC), training time, inference time, confusion matrix, feature importance, as well as One-Way Analysis of Variance (ANOVA) and post-hoc Tukey’s Honestly Significant Difference (Tukey HSD) tests. The experimental results show that CatBoost achieved the highest classification performance, with an Accuracy of 84.86%, an F1-Score of 87.19%, a Recall of 88.03%, and an ROC-AUC of 0.9220. XGBoost delivered comparable predictive performance with a substantially shorter training time (3.92 seconds), while LightGBM was the most computationally efficient algorithm, requiring an average training time of only 0.85 seconds. In contrast, the Stacking Ensemble required the longest training time (215.41 seconds) without providing a statistically meaningful performance improvement over the boosting-based models. The ANOVA results revealed significant differences in ROC-AUC among the evaluated models (p < 0.001), whereas the Tukey HSD test indicated that most boosting algorithms did not differ significantly from one another. These findings provide practical guidance for selecting classification algorithms by balancing predictive performance and computational efficiency to support early mental health screening systems for university students.
Abstract:Financial crisis is a serious threat to economic stability, so an effective early warning system is needed to detect and anticipate potential risks early on. The implementation of artificial intelligence (AI) in financial…
l crisis early warning systems offers significant advantages through big data analysis capabilities, non-linear pattern detection, and more accurate and faster risk prediction than conventional methods. Studies have shown that machine learning and deep learning algorithms can improve crisis prediction accuracy, expand the scope of risk monitoring, and support more responsive decision-making by regulators and financial industry players. However, challenges such as data quality, security, model transparency, and human resource readiness still need to be addressed for optimal AI implementation. This study concludes that with good governance, investment in infrastructure and human resources, and adaptive regulations, AI can be a strategic tool in strengthening early warning systems and maintaining financial sector resilience in the digital era.
Abstract:Artificial Intelligence has become a transformative technology in the field of financial technology, leveraging advanced algorithms and machine learning to identify risks and make informed decisions. However, its widespread…
ead adoption presents new challenges related to ethical use, data privacy, security concerns, potential bias, and discrimination. This study aims to explore the benefits of AI-based risk management in Fintech while highlighting associated challenges and providing recommendations. This research utilises the systematic review methodology to analyse existing literature and identify important patterns, gaps, and areas for further investigation. The study utilised data gathered from the Scopus database to obtain credible scholarly materials. Research data was collected from a variety of countries including the United States, China, European nations, and other Asian countries in order to develop a comprehensive understanding of AI-based risk management on a global scale. The findings highlight the crucial role of ethical considerations in implementing AI-based risk management systems to ensure fairness, transparency, and accountability. Moreover, the fintech industry needs to establish strong data protection measures and address issues related to bias and discrimination in order to instil trust and uphold public confidence in AI-based risk management. Future research should emphasise assessing the effectiveness of different algorithms and approaches while also examining potential regulatory frameworks and legal implications associated with AI-based risk management strategies.
Abstract:This research aims to understand the influence of Sales Promotion, Utilitarian, Self-Esteem, and Hedonic Motive on Purchase Decisions with Impulse Buying and Behavior Intention as intervening variables in Shopee. The research…
earch method used is quantitative with a descriptive approach using a Likert scale. Using the SEM data analysis method. The population of this study is Shopee application users in Indonesia, especially generation Z (1997-2012), with 285 respondents taken using non-probability sampling techniques. The research results show that Sales Promotion, Utilitarian, Self-Esteem, and Hedonic Motive have a positive effect on Purchase Decision. Apart from that, Sales Promotion also has a positive effect on Impulse Buying and Behavior Intention. The implication of this research is that companies can utilize the Self-Esteem motive to improve purchasing decisions. Personalized product recommendations and price offers that match consumer preferences can be implemented to increase consumer satisfaction and trust. Implementing algorithms can help companies provide better shopping experiences, improve purchasing decisions, and strengthen consumer behavioral intentions
Abstract:The rapid growth of the hotel industry requires hotels to improve operational planning, one of which is by forecasting room reservations. Inaccurate forecasting may cause an imbalance between room availability and customer…
er demand. This study aims to compare the Weighted Moving Average and Single Exponential Smoothing algorithms in forecasting room reservations at Raz Hotel and Convention Medan using historical data from January 2025 to May 2026. The research method consisted of data collection, forecasting using both algorithms, and accuracy evaluation through Mean Absolute Deviation, Mean Squared Error, and Mean Absolute Percentage Error. The results indicate that the Single Exponential Smoothing algorithm achieved a Mean Absolute Percentage Error of 15.24%, which is lower than the 15.63% obtained by the Weighted Moving Average algorithm. Furthermore, the Single Exponential Smoothing algorithm predicted 835.90 room reservations for June 2026. Therefore, it can be concluded that the Single Exponential Smoothing algorithm provides better forecasting accuracy and is more suitable for predicting room reservations at Raz Hotel and Convention Medan.
Abstract:This study aims to explore the impact of social media use on the mental health of university students. Utilizing a qualitative approach, this study analyzes data from in-depth interviews with 10 students from various universities…
versities in Indonesia. Thematic analysis reveals that social media use contributes to various aspects of mental health, including anxiety, depression, and self-image disorders. The study identifies that interactions on social media often trigger social comparison and unrealistic expectations about personal life, contributing to a decline in psychological well-being. However, the results also indicate that social media can be a useful tool for supporting mental health, particularly through support communities and awareness of mental health issues. The researchers suggest a need to increase awareness of healthy and constructive social media use among students. This study also emphasizes the importance of mental health education and intervention programs in educational institutions. These results provide insights for policymakers, educators, and mental health practitioners in developing effective strategies to support the mental health of students in the digital era. This study paves the way for further research on the specific impact of social media features and algorithms on student mental health.
Abstract:The development of digital technology over the past two decades has revolutionized human communication, social interaction, and the formation of both individual and collective identities in a global context. Digital media…
a no longer merely function as communication tools but have become socio-cultural spaces that actively shape values, norms, symbols, and everyday practices. This phenomenon raises research questions regarding how cultural transformation and identity construction are influenced by the dominance of digital technology, as well as how social media and platform algorithms shape cultural visibility and digital identities. This study employs a qualitative approach through a critical narrative literature review, synthesizing theories of network society, mediatization, cultural globalization, digital capitalism, and digital identity. The analysis focuses on four key issues: cultural digitalization, the role of social media in identity formation, the emergence of hybrid identities, and platform power in determining cultural representation. Findings indicate that digital technology has ambivalent effects: it expands spaces for cultural expression, enhances public participation, and provides opportunities for marginalized groups to represent their identities; however, global platform dominance drives cultural homogenization, identity commodification, and inequitable symbolic representation. The novelty of this study lies in integrating a critical perspective on the influence of algorithms and digital capitalism in identity formation, which has rarely been addressed comprehensively in digital cultural studies. These findings offer theoretical contributions to media and cultural research as well as practical implications for understanding identity representation in the global digital era.
Abstract:Stunting remains a crucial public health issue in Indonesia as it has multidimensional impacts on physical growth, cognitive development, productivity, and the long-term quality of human resources. This study aims to analyze…
lyze the potential application of predictive Artificial Intelligence (AI)-based Wearable Health Technology in preventing child stunting, with an emphasis on the integration of technology into the health system, socio-cultural interrelations, and ethical dimensions that need to be anticipated. The research employed a qualitative approach through a systematic literature review, analysis of national health policies, and examination of reports on the implementation of digital health technologies related to child health. The results indicate that wearable devices are effective in recording health indicators in real-time, while predictive AI algorithms enable early detection of stunting risks, allowing interventions to be carried out quickly, accurately, and in a personalized manner. However, the success of implementation is highly dependent on the support of the digital health ecosystem, including adaptive regulations, infrastructure readiness, digital literacy among communities, as well as multi-stakeholder collaboration involving the government, medical professionals, the technology industry, and families. This study affirms that the integration of Wearable Health Technology and predictive AI is not merely an additional tool in medical services, but rather a representation of a health system transformation towards a more preventive, predictive, precise, and participatory paradigm. Thus, the utilization of evidence-based digital innovations has the potential to strengthen the effectiveness of stunting prevention programs in Indonesia while accelerating the achievement of national health development targets within the context of technological disruption and global dynamics.
Abstract:Abstract: This study aims to analyze the influence of gender on user responses to social media strategies on the Instagram platform by utilizing data science techniques. The methodology includes collecting user interaction…
on data based on gender, statistical analysis, and applying machine learning algorithms to identify response patterns. The results reveal significant differences in how male and female users respond to social media content and campaigns, affecting marketing strategy effectiveness. Data science analysis showed that the K-Means Clustering method segmented users into three groups based on interaction patterns, with female users showing the highest engagement rate (18%) compared to male users (12%). The Decision Tree model identified gender as the most dominant predictor of engagement (40%), followed by user growth (25%) and content type (20%). The Random Forest model validated that gender-targeted strategies increased marketing effectiveness by up to 22%. Multivariate regression revealed a positive effect of female user proportion (+0.45) and a negative effect of male user proportion (−0.12) on engagement. In conclusion, a deeper understanding of gender-based response differences can assist companies in designing more targeted social media strategies and enhancing engagement.
Keywords: gender; social media response; instagram; data science; marketing strategy
Abstrak: Penelitian ini bertujuan untuk menganalisis bagaimana pengaruh gender memengaruhi respons pengguna terhadap strategi media sosial di platform Instagram dengan memanfaatkan teknik data sains. Metode yang digunakan meliputi pengumpulan data interaksi pengguna berdasarkan gender, analisis statistik, dan penerapan algoritma machine learning untuk mengidentifikasi pola respons. Hasil penelitian menunjukkan adanya perbedaan signifikan dalam cara pengguna laki-laki dan perempuan merespons konten dan kampanye media sosial, yang berdampak pada efektivitas strategi pemasaran. Analisis data sains menunjukkan bahwa metode K-Means Clustering berhasil mengelompokkan pengguna ke dalam tiga segmen berdasarkan pola interaksi, dengan segmen perempuan menunjukkan tingkat engagement tertinggi (18%) dibandingkan laki-laki (12%). Model Decision Tree mengidentifikasi gender sebagai faktor paling dominan terhadap engagement dengan kontribusi sebesar 40%, disusul oleh pertumbuhan pengguna (25%) dan jenis konten (20%). Random Forest memvalidasi bahwa strategi yang disesuaikan berdasarkan gender meningkatkan efektivitas pemasaran hingga 22%. Regresi multivariat menunjukkan bahwa proporsi pengguna perempuan berkontribusi positif sebesar 0,45 poin terhadap engagement, sedangkan pengguna laki-laki berkontribusi negatif sebesar -0,12. Kesimpulannya, pemahaman mendalam tentang perbedaan respons berdasarkan gender dapat membantu perusahaan dalam merancang strategi media sosial yang lebih tepat sasaran dan meningkatkan engagement.
Kata kunci: gender; respons media sosial; instagram; data sains; strategi pemasaran