This study aims to compare raw and functional data within a decision tree framework. The study examines the patterns exhibited by countries across various variables based on the number of deaths per million during the COVID-19 pandemic. Vectorial inputs are transformed into functional inputs, and a smoothed dataset representing the raw data with fewer functional observations is obtained. Regression and classification trees obtained from smoothed and unsmoothed data are then compared. Additionally, changes in the decision tree model are examined with the functions obtained by changing the number of basis functions. Using the functional data approach, the importance of variables affecting mortality rates is identified, and changes in prediction results arising from the transformation of the data structure from raw to functional form are analyzed. The main contribution of the study is the implementation of decision tree analyses using functional objects. The findings indicate that fitting the data through a basis function approach provides higher classification performance than using raw data. Moreover, functional data–based trees achieve comparable or higher accuracy with simpler and more interpretable tree structures. Based on these findings, classification decision tree analysis using smoothed data is proposed as an alternative approach for the analysis of functional predictors.
Today, the retail sector is in a state of constant transformation. Sales promotion activities play a critical role in an environment where it is increasingly difficult for brands to boost their sales and strengthen their market positions. Insert campaigns are one of the most common and traditional methods among these activities, which can directly influence consumers’ purchasing decisions. This study aims to reveal the distribution of periodic campaigns by year, the product variety within these campaigns, the status of product prices before and after the campaigns, the sales volumes and rates of products included in the campaigns, and whether the campaigns affect the retail sector’s overall sales volume. For this purpose, the campaign performances of 125 products from 6 main categories and 31 subcategories included in 13 two-day insert campaigns conducted in 2024 were analyzed using data obtained from 10 local supermarket branches. The distribution of campaigns throughout the year can be considered an indication that the retail sector strives to maintain active year-round communication with customers. Although there is general product variety, it was observed that campaigns conducted especially in the initial and final months of the year included a wider range of products compared to others. The emphasis on products perceived as essential, ongoing, and non-deferrable needs in campaigns can be interpreted as an attempt to attract consumer interest, ensure customer loyalty by meeting basic needs, and increase store traffic to gain a competitive edge. It was found that prices declined during campaign days, discounts of varying rates were applied across different campaigns and product groups, and the retail sector used attractive price reductions as a strategic tool. Sales volumes increased on campaign days, and campaigns played a significant role in boosting sales; however, this effect varied depending on the timing of the campaign, the applied discount rate, and the products’ main categories and subcategories. It was determined that consumers concentrated on their shopping during campaigns offering high discounts and tended to wait for such campaigns to purchase products that are more stockable and relatively flexible in consumption frequency. The findings and results of this study provide important, valuable, and critical insights for businesses to improve their campaign effectiveness and strategies.
Intelligent Transportation Systems (ITS) are central to making urban public transport more efficient, sustainable, and equitable. This study evaluates the operational performance and spatial configuration of nine urban bus lines in the Bandırma district of Türkiye by jointly analyzing daily passenger records and route geometry. A total of 1,265 daily observations, representing approximately 2.23 million passenger movements over the March–July 2022 period, are examined, and actual route lengths are derived from coordinate sequences using the haversine formula. Four key performance indicators—average daily ridership, passengers-per-kilometer efficiency, the weekend-to-weekday demand ratio, and the coefficient of variation—are integrated through the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), a multi-criteria decision-making method, to rank the lines objectively. Spatial behavior is characterized through a route-density map overlaid on a CARTO Positron basemap, which exposes service overlaps and coverage gaps, while inter-line demand co-movement is quantified using Pearson correlation. The results indicate that Line 10 (Faculty–Hospital) attains the highest performance (TOPSIS score 0.886; 5,398 passengers/day; 568.8 passengers/km), whereas Line 9 (Livatya) ranks lowest (0.004). A mean inter-line correlation of 0.63 reveals a system-wide common demand pattern, implying that scheduling decisions should be coordinated across lines. The proposed framework offers a reproducible, data-driven decision-support tool for trip planning, resource allocation, and route optimization in small- and medium-sized cities.Intelligent Transportation Systems (ITS) are central to making urban public transport more efficient, sustainable, and equitable. This study evaluates the operational performance and spatial configuration of nine urban bus lines in the Bandırma district of Türkiye by jointly analyzing daily passenger records and route geometry. A total of 1,265 daily observations, representing approximately 2.23 million passenger movements over the March–July 2022 period, are examined, and actual route lengths are derived from coordinate sequences using the haversine formula. Four key performance indicators—average daily ridership, passengers-per-kilometer efficiency, the weekend-to-weekday demand ratio, and the coefficient of variation—are integrated through the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), a multi-criteria decision-making method, to rank the lines objectively. Spatial behavior is characterized through a route-density map overlaid on a CARTO Positron basemap, which exposes service overlaps and coverage gaps, while inter-line demand co-movement is quantified using Pearson correlation. The results indicate that Line 10 (Faculty–Hospital) attains the highest performance (TOPSIS score 0.886; 5,398 passengers/day; 568.8 passengers/km), whereas Line 9 (Livatya) ranks lowest (0.004). A mean inter-line correlation of 0.63 reveals a system-wide common demand pattern, implying that scheduling decisions should be coordinated across lines. The proposed framework offers a reproducible, data-driven decision-support tool for trip planning, resource allocation, and route optimization in small- and medium-sized cities.
Polarization on social media has become a prominent topic in computational social science, as it shapes public discourse, intensifies societal divisions, and influences collective dynamics. Despite the rapid growth of computational approaches, there remains limited consensus regarding how polarization is operationalized, measured, and validated across different analytical frameworks. This study aims to systematically examine the methodological landscape of computational polarization research by analyzing the conceptualization, quantification, and validation of polarization in empirical social media studies. A PRISMA-guided systematic literature review of 70 empirical studies was conducted, retrieved from the Web of Science Core Collection. A structured coding framework, including analytical paradigms, operational definitions, measurement techniques, data sources, and validation strategies. Using this analytical framework, the reviewed studies were examined across three dominant methodological paradigms: network-based, content-based, and hybrid approaches. The analysis demonstrates that these paradigms rely on distinct assumptions, analytical units, and measurement strategies. Network-based approaches primarily capture structural separation through interaction patterns and community structures, whereas content-based approaches examine semantic, affective, and ideological divergence through textual signals. Hybrid approaches integrate relational and content information to capture the complementary dimensions of polarization. However, the findings indicate that these approaches capture different manifestations of polarization rather than measuring a unified construct. Furthermore, the review highlights persistent challenges, including platform-specific data biases, limited temporal analysis, heterogeneous validation practices, and the lack of standardized benchmarks for comparing polarization measures. This review advances understanding of the strengths, limitations, and applicability of computational polarization approaches and highlights future directions for developing more comprehensive, validated, and comparable measurement frameworks.
As the integration of artificial intelligence (AI)–based educational technologies into teaching and learning processes accelerates, trust has emerged as a key determinant of whether preservice teachers are willing to adopt these technologies. The present study aimed to examine the relationship between the perceived trust of pre-service science teachers in AI-based educational technologies and generative AI acceptance and to determine the predictive power of trust on acceptance. Grounded in the TAM, the study employed a cross-sectional correlational survey design to investigate the role of trust in acceptance processes. Data were collected from 270 pre-service science teachers enrolled in a science teacher education program at a university via an online questionnaire. The Trust in AI-Based Educational Technologies Scale (TAI-EdTech Trust) and the Generative AI Acceptance Scale (GAIA) were used for data collection. The findings revealed a moderate, positive, and statistically significant relationship between trust in AI-based educational technologies and GA acceptance, indicating that trust significantly predicted GA acceptance. The results highlight the importance of developing content in science teacher education programs that supports the safe, ethical, and pedagogically responsible use of generative AI tools.
