PURPOSE This study optimized stopword removal to enhance topic modeling performance. We propose an objective method combining normalized pointwise mutual information (NPMI) with median-based term frequency–inverse document frequency (TF–IDF) to automatically generate stopwords. METHODS Using text data from 443 research papers on “Taekwondo sparring,” we selected stopword candidates based on NPMI and identified 30 words with the lowest TF–IDF scores. We examined the impact of removing 1–30 stopwords on u_mass coherence scores. RESULTS The NPMI–TF–IDF method significantly improved coherence (R² = .456; p < .001). However, excessive removal led to diminishing returns, with the optimal coherence score (−11.442) achieved at 200 stopwords. In contrast, manually selected stopwords yielded a lower coherence score (−16.001). The findings indicate that integrating TF–IDF with NPMI effectively preserves meaningful words and outperforms PMI2 and PMI3 approaches. CONCLUSIONS Manual stopword selection can reduce reproducibility. Optimizing stopword removal based on domain-specific characteristics is essential. Future research should validate this method across diverse fields to establish a more generalizable standard.
PURPOSE This study aims to verify the reliability of a newly developed augmented reality (AR)-based agility measurement method by comparing its outcomes with those of conventional agility tests. METHODS The participants were 12 middle and high school weightlifting athletes. To assess the reliability between the initial field test and the subsequent AR-based test, a dependent t-test and correlation analysis were conducted for three variables: side-step (count), color catch (count), and color catch (average reaction time). RESULTS First, significant correlations were observed between the field and AR-based measurements for side-step (count), color catch (count), and color catch (reaction time), confirming the reliability of the AR-based method. Second, no significant difference was found between side-step (count) and AR side-step (count) (t = 1.910), indicating consistency between the two methods. Third, a significant difference was found between color catch (count) and AR color catch (count) (t = 4.750), suggesting lower consistency in this variable. Fourth, no significant difference was observed between color catch (average reaction time) and AR color catch (average reaction time) (t = 2.100), and the significant correlation coefficient (.579) further supports their consistency. CONCLUSIONS These results suggest that the newly developed AR-based agility measurement method yields reliable data and has practical applicability for sports performance assessment and training in field settings.
PURPOSE This study assessed elite Taekwondo athletes’ physical fitness and developed reference standards for both their basic and specialized physical fitness. METHODS Data for analysis were collected from 870 athletes: from national teams, 123 elite Taekwondo athletes from the Performance Analysis and Assessment System (PAAS) administrator website (1999–2020); from regional sports centers, 731 collegiate and general division elite Taekwondo athletes (2015–2019); and from Y University, 16 elite Taekwondo athletes. Through measurement items’ selection and categorization, 20 physical fitness items were selected for the reference standards’ development, including 9 for basic fitness and 11 for specialized fitness. Taekwondo weight classes were divided into two: light + middle (fin, fly, bantam, feather) and middle + heavy (light, welter, middle, heavy). RESULTS Descriptive statistics for basic and specialized physical fitness items were categorized by gender and athletes’ fitness level. The reference standards’ development was aligned with existing standards, integrating the Cajori physical fitness 5-levels. It also introduced minimum physical fitness reference standards and target achievement reference standards for evaluating elite Taekwondo athletes’ physical fitness. CONCLUSIONS The reference standards proposed here can serve as objective indicators in selection of national representative athletes and also provide foundational data to establish fitness goals and evaluate future elite athletes’ physical fitness.
Purpose The purpose of this study was to find out the physical characteristics of Wushu athletes by comparing the differences on the results of physical fitness between the male athletes of the Wushu national team Taolu and the Sanda group. Methods Measurement of basic and professional fitness based on muscle function, targeting 37 men Wushu national team players (24 taolu, 13 Sanda) in the selection and evaluation contests twice in 2018 and 2019. Body composition, isometric muscle strength, flexibility and equilibrium, anaerobic power, and isokinetic muscle strength. The fitness factors were divided into two groups, Taolu and Sanda. Results First, body fat rate of the Sanda athlete group was significantly lower than that of the taolu athlete group (p<.01). Second, in the isometric muscle strength category(back muscle strength, grip strength), the Sanda athlete group had higher muscle strength than Taolu athlete group, but there was a statistically significant difference only in the left grip strength (p<.01). Third, in terms of flexibility and equilibrium, the taolu players were significantly higher in all items(p<.001). Fourth, in the anaerobic power, the taolu athlete group had higher both the peak power and the mean power, and there were a significant differences(peak power: p<.01, mean power: p<.001). Fifth, isokinetic muscle strength was significantly higher in the right knee flexion of the taolu athlete group (p<.01), and lumbar extensor muscle was significantly higher in the Sanda athlete group (p<.05). Sixth, in the isokinetic strength ratio, the knee flexion ratio of the Sanda athlete group were significantly higher on the left and right knee flexion and extensor ratios (p<.05). In addition, in the lumbar flexor and extensor ratios, the group of Sanda athlete group were significantly higher on the lumbar extensor(p<.05). Seventh, there was no significant difference between two groups in isokinetic muscle power. Conclusions The results of this study can be used as basic data to improve the efficiency of technical and physical training through the analysis of the characteristics of Taolu and Sanda. The effectiveness of this training will help to improve the performance.
PURPOSE This study developed a video-based taekwondo match-analysis program and validated its usefulness as a match-analysis tool; women’s match videos from international competitions held after the revision of competition rules in June 2022 were used for this. METHODS Analysis was conducted on 78 women’s matches from international competitions that were held after the June 2022 revision of the rules, including the Chuncheon Asian Taekwondo Championships and the Rome Grand Prix. The analysis was performed using IBM SPSS Statistics 26 and jamovi 1.1.9.0 to examine the independence and conditional efficiency of linked techniques, perform cross-tabulation analyses, and conduct hierarchical log-linear analyses. The odds ratios of key techniques with reference to match outcomes were calculated using Microsoft Excel. The match analysis program was developed in a Python 3.x environment. RESULTS The video-based taekwondo match-analysis program was shown to be an effective tool for match analysis, integrating technique tagging, frame-synchronized analysis, collection of athlete position coordinates, and scoring and statistical functions that reflected both pre- and post-rule revisions. In addition, weight class, attack type, and key technique execution were identified as significant factors that explained match outcomes. Among the key techniques, punches was a relatively higher execution frequency and odds ratio in winning matches. CONCLUSIONS The video-based taekwondo match-analysis program that was developed in this study demonstrates practical usefulness as an empirical performance tool of analysis capable of integratively collecting and analyzing technical execution, competition rules, spatial information, and scoring data.
