Earticle

Much of modern medical research are supported by statistics and machine learning techniques. For the reliability of these studies, it is important to measure the effectiveness of the developed models in making predictions on real medical data. Cross validation is a tool, which provides a principled method of measuring the effectiveness of models and comparing models with each other. This paper presents theoretical and empirical analysis of common cross validation methods. Using sample logistic regression model along with with six different validation methods, comparative analysis has been conducted. The model was developed to determine the level of thyroglobulin hormone, thyroid cancer marker. The marker indicates an increased risk of thyroid cancer metastasis, and is commonly used in metastasis diagnosis. Whereas, literature does not provide a single threshold of thyroglobulin value, indicating an increased risk of metastasis. This paper confirms that one of the main reasons of this lack, is the high variance of the developed models. This research argues that the use of the validation technique also influences both: measures of the model quality and the threshold value. The results show a high discrepancy in the determination of threshold values. However, iterative methods (such as sampling and bootstrap) seem to produce more stable outcomes.

Purpose: To develop and validate the short form of the Korean adaptation of the Cultural Competence Scale for Nurses. Methods: To shorten the 33-item Cultural Competence Scale for Nurses, an expert panel (N ¼ 6) evaluated its content validity. The revised items were pilot tested using a sample of nine nurses, and clarity was assessed through cognitive interviews with respondents. The original instrument was shortened and validated through item analysis, exploratory factor analysis, convergent validity, and reliability using data from 277 hospital nurses. The 14-item final version was cross-validated through confirmatory factor analysis, convergent validity, discriminant validity, known-group comparisons, and reliability using data from 365 nurses belonging to 19 hospitals. Results: A 4-factor, 14-item model demonstrated satisfactory fit with significant factor loadings. The convergent validity between the developed tool and transcultural self-efficacy was significant (r ¼ .55, p < .001). The convergent validity evaluated using the Average Variance Extracted and discriminant validity were acceptable. Known-group comparisons revealed significant differences in the mean scores of the groups who spent more than one month abroad (p ¼ .002) were able to communicate in a foreign language (p < .001) and had education to care for foreign patients (p ¼ .039). Cronbach's a was .89, and the reliability of the subscales ranged from .74 to .91. Conclusion: The Cultural Competence Scale for Nurses-Short Form demonstrated good reliability and validity. It is a short and appropriate instrument for use in clinical and research settings to assess nurses' cultural competence.

Purpose: The purpose of this study was to culturally adapt the Intensive Care Experience Questionnaire (ICEQ) for Korean patients and evaluate its validity and reliability. Methods: This cross-sectional study comprised two phases. The first phase involved the cultural adaptation of the ICEQ. In the second phase, the psychometric properties of the adapted measure were evaluated. Cultural adaptation was conducted in accordance with the World Health Organization's process. The adapted questionnaire was administered to 200 Korean patients who had received treatment in an intensive care unit within the past six months. Model fit was assessed through confirmatory factor analysis, and convergent validity and discriminant validity of the items were assessed. Knowngroups validity was evaluated using the t test and Cohen's effect size. Cronbach's a was used to examine internal consistency reliability. Results: The Korean version of the ICEQ (K-ICEQ) consists of 26 items and four subscales: Frightening Experiences, Awareness of Surroundings, Satisfaction with Care, and Recall of Experiences. The model fit indices, convergent validity, and discriminant validity of the K-ICEQ were all satisfactory. According to the results of the test of known-groups validity, intensive care unit (ICU) experience varied according to gender, planned ICU admission, mechanical ventilation, and restraints application. Cronbach's a of the KICEQ subscales ranged from .74 to .93. Conclusion: The validity and reliability of the K-ICEQ reflecting the characteristics of Korean ICU patients were satisfactory. The K-ICEQ can be used to evaluate the experience of Korean ICU patients and contribute to the development of interventions to improve the ICU experience.

Objective To predict the risk of falls, Fall Risk Assessment (FRA) system has been newly developed to measure multi-systemic balance control among community-dwelling older adults. The aim of this study was to examine the association between FRA and fall-related physical performance tests. Methods A total of 289 community-dwelling adults aged 65 years and older participated in this cross-sectional study. All participants underwent FRA test and physical performance tests such as Short Physical Performance Battery (SPPB), Berg Balance Scale (BBS), and Timed Up and Go Test (TUG). Results Participants who were younger, male, highly educated, living with family members, having high body mass index, having high appendicular lean mass index, and having no irritative lower urinary tract syndrome were more likely to have higher FRA scores. SPPB (β=1.012), BBS (β=0.481), and TUG (β=-0.831) were significantly associated with FRA score after adjusting for the variables (all p<0.001). Conclusion FRA composite score was closely correlated with SPPB, BBS, and TUG, suggesting that FRA is a promising candidate as a screening tool to predict falls among community-dwelling elderly people.

Purpose: Patients with atrial fibrillation (AF) experience information overload during self-managementand engagement in health care. The Atrial Fibrillation Information Overload Scale (AFIOS) has beenvalidated and used to assess information overload in Australian AF population. However, a validatedChinese version of AFIOS is not available. The aim of this study was to translate the 8-item and modified 5-item AFIOSs from English into Chinese and test their psychometric properties. Methods: Translation and cross-cultural adaption were conducted according to international guidelines. The translated scales were administered to 362 patients with AF from May 2022 to March 2023. Psychometricevaluation included item analysis, content validity, construct validity, and reliability. Results: All items had item-total correlation coefficient over .30 and were significantly different betweenhigh-scoring and low-scoring groups. Exploratory factor analysis revealed that both versionswere one-factor structure, explaining 52.1% and 52.7% of the total variance. Confirmatory factor analysisindicated that the one-factor model fit the data well in both versions. The scale level-content validityindex, Cronbach α, and test?retest reliability of the 8-item scale were .96, .87, and .97, respectively;those of the 5-item scale were .97, .75, and .95, respectively. Conclusion: Chinese versions of both the scales exhibit good psychometric properties. Healthcare professionalscould utilize the 5-item version scale to assess AF patients’ information overload and developtargeted strategies to mitigate its negative impact on self-management and engagement.

Objective: To culturally adapt the original English Scale for the Assessment and Rating of Ataxia to Thai (SARA-TH) and to evaluate the reliability and validity of the SARA-TH in assessing ataxia in acute ischemic stroke or transient ischemic attack (TIA) patients, as assessed by three healthcare professionals.Methods: The SARA underwent translation and cross-cultural adaptation to Thai according to established guidelines. Reliability (e.g., internal consistency, intrarater reliability, interrater reliability) and validity (e.g., content validity, convergent validity) were assessed in a sample of 50 participants with ataxia after acute ischemic stroke or TIA. Spearman correlation analysis was used to examine the relationships between the SARA-TH and the Barthel Index (BI-TH), the National Institutes of Health Stroke Scale (NIHSS-TH), and the International Cooperative Ataxia Rating Scale (ICARS) to assess convergent validity. Interrater and intrarater reliability among experienced and novice neurologists, physiotherapists, and occupational therapists were assessed using weighted kappa.Results: The SARA-TH demonstrated good comprehension and exhibited no significant floor or ceiling effects. It showed excellent internal consistency (Cronbach’s α≥0.776). Significant correlations were found between the SARA-TH score and the BI-TH score (rs=-0.743 to -0.665), NIHSS- TH score (rs=0.404?0.513), and ICARS score (rs=0.859?0.917). The intrarater reliability for each rater ranged from 0.724 to 1.000 (p<0.01), and the interrater reliability varied from 0.281 to 0.927 (p<0.01).Conclusion: The SARA-TH has excellent internal consistency, validity, and intrarater reliability, as well as acceptable interrater reliability among health professionals with varying levels of experience. It is recommended for assessing ataxia severity in individuals following acute ischemic stroke or TIA.

Purpose Screening for hepatocellular carcinoma (HCC) as an effective instrument to reduce the burden of late diagnoses remains underutilized in China, much of the Asian countries, and in a sense all over the world. Modifiable psychosocial factors should be identified to improve screening utilization and reduce the burden of late diagnoses. However, valid psychosocial measures are unavailable. This study aimed to translate, culturally adapt, and validate the preventive health model (PHM) instrument for measuring psychosocial factors of HCC screening among patients with hepatitis B. Methods This study was conducted from June 2020 to April 2021 in three rigorous phases: (1) committee-based translation from English to Chinese; (2) cognitive interviews (n = 33) and Delphi expert consultations (n = 7) for cultural adaptation; and (3) a cross-sectional study (n = 305) for validation. Results In phase I, two items were reworded, and two retranslated for semantic equivalence. In phase II, issues related to comprehension, sensitive wording, wording clarity, question relevance, and cultural sensitivity were addressed by including pictures, rewording five items, and developing seven items. In phase III, exploratory and confirmatory factor analyses suggested a five-factor 20-item solution: it explained 76.9% of the variance; had adequate factor loading (.60?.91), convergent and discriminant validity; satisfactory model fit indices; and reliability (Cronbach's α, .86?.91). Known-group analysis showed that patients with optimal HCC screening behavior had significantly higher scores on each subscale than those not having such. Conclusion The Chinese PHM instrument is culturally sensitive, reliable, and valid to measure the psychosocial factors of HCC screening. It can help nurses and researchers to tailor strategies to improve clinical HCC screening practices in high-risk HCC regions.

Objective: To translate, culturally adapt, and validate the Roland-Morris Disability Questionnaire (RMDQ) for Malay-speaking individuals with low back pain (LBP) in Malaysia.Methods: A methodological study was conducted following established international guidelines for cross-cultural adaptation and psychometric validation. The process included forward?backward translation, expert review, content validation, face validation, and psychometric testing among individuals with LBP. Content and face validity indices were evaluated by rehabilitation professionals (n=8) and individuals with LBP (n=9). Psychometric testing was conducted with 102 participants to assess internal consistency, test-retest reliability (n=31), concurrent validity using the Malay Oswestry Disability Index, and known-groups validity.Results: All items demonstrated excellent content and face validity (I-CVI, I-FVI, S-CVI/Ave, and S-FVI/Ave=1.00). The Malay RMDQ showed high internal consistency (Cronbach’s α=0.932) and excellent test-retest reliability (ICC=0.996). Concurrent validity with the Malay Oswestry Disability Index was moderate (r=0.742, p<0.001). A preliminary known-groups comparison showed a statistically significant difference in the Malay RMDQ scores between participants with LBP and those without LBP (p=0.03).Conclusion: The Malay RMDQ demonstrated preliminary psychometric support, including evidence of reliability, validity, and cultural appropriateness, for assessing disability in individuals with LBP. Future research should explore its responsiveness and utility in broader populations.

Objective To perform a translation and cross-cultural adaptation of the Cardiac Rehabilitation Barriers Scale (CRBS) for use in Korea, followed by psychometric validation. The CRBS was developed to assess patients’ perception of the degree to which patient, provider and health system-level barriers affect their cardiac rehabilitation (CR) participation.Methods The CRBS consists of 21 items (barriers to adherence) rated on a 5-point Likert scale. The first phase was to translate and cross-culturally adapt the CRBS to the Korean language. After back-translation, both versions were reviewed by a committee. The face validity was assessed in a sample of Korean patients (n=53) with history of acute myocardial infarction that did not participate in CR through semi-structured interviews. The second phase was to assess the construct and criterion validity of the Korean translation as well as internal reliability, through administration of the translated version in 104 patients, principle component analysis with varimax rotation and cross-referencing against CR use, respectively.Results The length, readability, and clarity of the questionnaire were rated well, demonstrating face validity. Analysis revealed a six-factor solution, demonstrating construct validity. Cronbach’s alpha was greater than 0.65. Barriers rated highest included not knowing about CR and not being contacted by a program. The mean CRBS score was significantly higher among non-attendees (2.71±0.26) than CR attendees (2.51±0.18) (p<0.01).Conclusion The Korean version of CRBS has demonstrated face, content and criterion validity, suggesting it may be useful for assessing barriers to CR utilization in Korea.

.

Purpose: This study aimed to validate the Child Healthy Lifestyle Profile as an instrument for screening healthy behaviors in school-aged children in South Korea. Methods: Self-reported questionnaires were administered to 454 students, comprising elementary-school students (n=221) and child cancer survivors (n=233). Reliability and validity were assessed using Cronbach's ?, exploratory factor analysis (EFA), and confirmatory factor analysis (CFA). Results: Cronbach's ?, as a reliability test, was 0.87. Varimax rotation yielded nine factors with eigenvalues greater than 1 in the EFA, which explained 61.0% of the total variance. In the CFA, both convergent and discriminant validities were acceptable. Therefore, the Child Healthy Lifestyle Profile was validated as an assessment tool for Korean school-aged children. Conclusion: Based on the results of this study, the Child Healthy Lifestyle Profile was identified as a reliable and valid instrument for assessing healthy lifestyles in elementary-school children in South Korea.

Purpose: This study aimed to develop and compare machine learning models for predicting prediabetes in young adults in Korea using dietary intake data and to identify the most effective model. Methods: Data from the ninth Korea National Health and Nutrition Examination Survey were used, with 823 participants aged 19-35 years selected after excluding those with missing data. Logistic regression, k-nearest neighbors, and random forest models were applied to predict prediabetes, and the analysis was conducted using the Orange 3.5 program. Five-fold cross-validation was performed to reduce performance variability, and test data were used for final model validation. Results: In the dataset, 14%-15% of participants were classified as having prediabetes. The random forest model showed the highest performance in terms of classification accuracy, harmonic mean of precision and recall, and precision. Logistic regression had the highest performance regarding the model’s ability to distinguish between individuals with and without prediabetes. Age, thiamine intake, and water intake emerged as the most important predictors. Conclusion: This study demonstrated the utility of using dietary intake data to predict prediabetes in young adults. The random forest model provided the highest prediction accuracy, supporting early detection and intervention, which could help to reduce unnecessary treatment. This highlights nurses’ important role in educating patients about lifestyle changes and implementing preventive care. Future studies should incorporate additional factors, such as psychological and lifestyle variables, to improve the model's performance.

본 연구는 한국과 뉴질랜드환경에서 시장점유율에 따른 브랜드들의 이중위험에 대한 주제로 작은 시장점유율을 가진 브랜드인 경우에는 그렇지 않은 경우보다 단위당 낮은 판매가격을 유지하게 됨으로 판매량이 감소되고 수익이 줄어드는 현상을 뉴질랜드와 비교 연구하였다. 14개의 FMCG 제품군을 중심으로 진행된 연구결과 고 시장점유율을 통한 높은 이익 전략은 서구권의 국가인 뉴질랜드보다 아시아권 국가인 한국에서 여전히 유효한 것으로 나타나며, 특히 점유율이 높은 시장에서 1위 브랜드인 경우에는 약 15%의 가격우위를 점하는 것으로 나타난다. 하지만 이에 반해 뉴질 랜드의 경우에는 그렇지 않은 것으로 나타난다.

This study empirically investigates the double jeopardy phenomenon in a Korean and a New Zealand context. The double jeopardy is that companies with a small market share tend to suffer not only smaller sales volumes but also suffer a lower price than the market leader. The research reported here analyses price and market share data for 14 categories of household goods in Korea and a smaller number in New Zealand. Analysis shows that, in Korea, leading brands do enjoy a price premium as predicted, of around 15%, but that there is little or no evidence of double jeopardy occurring in New Zealand. Based on this study, evidence suggests that market share is a strong valid strategic objective in the East.

세계시민주의는 지역주의의 반대되는 개념으로 1차원적인 구조로 오래전부터 고려되어져 왔었다. 본 연구에서는 다국적 샘플을 사용하여 기존의 연구에서 제안된 세계시민주의의 2차원적 이론의 타당성을 검증하였다. 매우 다른 문화를 가진 4개 국가 (미국, 한국, 터키, 우크라이나)에서 수집된 다국적 샘플을 사용하여 교차문화적 연구와 다국적 소비자 세분화 작업에 세계시민주의 척도의 사용 및 적용 가능성을 토론하고자 한다. 본 연구의 결론은 세계시민주의 와 지역주의에 대해 동일한 척도를 사용하여 측정하는 단일 구조의 2분법을 제안한다. 본 연구의 결과는 원래 기존의 연구에서 발견된 세계시민주의 와 지역주의 사이의 차이는 각 국가집단에서도 발견되었지만, 미국 샘플의 경우는 약하게 나타났다. 척도의 외적 타당성 검증에서, 세계시민주의는 범세계적 매체에 때한 노출, 해외 거주 유무, 거주형태, 교육 수준, 연령 등과 같은 변인에 밀접한 관련성을 가지는 것으로 드러났다. 전문 경영인 관점에서 세계시민주의에 대한 논의를 하였다.

Cosmopolitanism has traditionally been considered a unidimensional construct that is the opposite of localism. In this study, the authors use a multinational sample to evaluate the validity of the two-dimensional theory of cosmopolitanism proposed by previous study. Using a cross-national sample of respondents from four maximally dissimilar cultures—the United States, South Korea, Turkey, and Ukraine—the authors attempt to assess the extent to which the scale is valid as a cross-cultural instrument, thus facilitating its use in cross-national segmentation. The study provides support for the proposition that cosmopolitanism and localism are indeed two dimensions of a single construct, and that they may be measured using the same scale. The results show that the distinction between cosmopolitanism and localism found in the past study was true in every case, albeit with weaker results for the American dataset. The authors also find, as a result of testing the scale’s external validity, that cosmopolitanism has significant relationships with the following consumer variables: exposure to international media, experiences living abroad, the sizes and settings of dwellings, education level, and age. Following presentation of the results, the implications and limitations are discussed.

본 연구에서는 기존에 개발된 조직내/외 경력이동 모형(오인수, 서용원, 신강현, 2004)에 대한 확인적 연구로서 독립적 표본을 이용한 교차타당화 분석을 실시하였다. 또한 이러한 분석 과정을 거쳐 타당화된 당해 모형이 경력연구의 주요 직무관련 인구통계학적 변인인 직급과 직종에 따라 상이한 양상과 구조적 관계를 갖는지 검증하기 위하여 중다표본 분석을 추가로 수행하였다. 먼저 [연구 1]에서는 교차타당화를 위하여 대기업 사무/기술직 근로자 585명으로부터 수집된 자료를 분석하였는데 당해 모형을 성공적으로 재검증할 수 있었다. 또한 당해 모형에 조직외 경력이동 대신에 (단순)이직의도를 투입한 모형도 우수한 부합도 지수를 보여주었다. 더욱이 두 개념간의 영차 상관이 .86으로 매우 높았는데, 이는 실제적으로 두 개념이 근로자들에게 변별되지 못하고 있음을 나타낸다. 다음으로 [연구 2]에서는 직급(대리/사원급과 과/차장급)과 직종(영업직과 비영업직)에 따라 표본을 두 집단으로 나누어 조직내/외 경력이동모형에 대한 일반화 가능성을 알아보기 위하여 중다표본 분석을 실시하였다. 그 결과, 요인계수는 직급간 동일하였으나, 구조적 관계 측면에서는 직무만족이 조직내 경력이동의도에 미치는 영향 및 경력만족이 조직외 경력이동의도(혹은 이직의도)에 미치는 영향 등 두 가지 경로계수에서 사원/대리급이 과/차장급보다 더 큰 부적인 경로계수를 보였다. 하지만, 직종간에는 요인계수 및 경로계수를 포함한 모든 모수가 유의한 차이를 보이지 않았고, 따라서 당해 모형의 직종간 완전한 일반화가 확인되었다. 이는 직종별 경력관리보다도 직급별로 차별화된 경력관리를 하는 것이 더욱 효과적일 수 있음을 시사한다. 추가적으로 본 연구의 제한점, 미래 연구방향, 그리고 인적자원 관리 측면의 시사점이 논의되었다.

This study was intended to cross-validate and generalize the ‘internal/external career movement model’ proposed by Oh, Suh, and Shin (2004). To achieve these purposes, the study was conducted by taking the following steps. After cross-validation with an independent sample was carried out, the model was tested using multi-sample analysis to examine if the cross-validated model generalizes across two demographic variables such as job levels/positions (manager-level vs. non-managerial level) and job types (sales vs. non-sales). First, the ‘internal/external career movement model’ was cross-validated using 585 clerical employees working for a large Korean company. The result revealed that the model showed satisfactory fit indices, and thus it was successfully cross-validated. In addition, the alternative model inputing ‘turnover intention’ instead of ‘external career movement intention’ also indicated satisfactory fit indices. Moreover, the zero-order correlation between two variables was very high (r=.86), which suggested that employees could not distinguish those two measures empirically. Next, multi-sample analysis was conducted and it showed that a non-managerial (low job-level) sub-sample showed higher negative path coefficient than a managerial (high job-level) sub-sample in terms of the relationship between ‘job satisfaction’ and ‘internal career movement intention’ and that between ‘career satisfaction’ and ‘external career movement intention.’ However, there was no difference between sales and non-sales sub-samples in terms of the ‘internal/external career movement model’, so the model perfectly generalized. This suggested that the direction of career development and management practices should be based upon job levels rather than job types in Korea. In addition, research limitations, future research directions, and implications for human resource management were discussed.

Happy-QoL 척도는 정신분열병 환자의 삶의 질(Quality of Life, QoL)을 평가하기 위해 개발된 척도이다. 탐색적 요인분석에서 이 척도는 5개 요인(경제적 영역, 일반적 영역, 신체적 영역, 직업/사회적 영역, 정서적 영역)으로 구성되어 있었다. 본 연구에서는 Happy-QoL 척도의 요인모델이 새롭게 표집된 정신분열병 환자들의 자료에도 잘 부합되는 지를 확인하고자 하였다. 또한 상이한 시간간격에서 QoL 척도에 대한 정신분열병 환자의 반응이 시간적 안정성이 있는 지도 알아보았다. 연구자가 제안한 모델은 5개 요인간에 상호 상관이 있는 모델이었으며, 대안모델은 요인간 상관이 없는 모델 및 요인 1(경제적 영역)과 요인 4(직업/사회적 영역), 요인 2(일반적 영역)와 요인 3(신체적 영역)간의 상관을 가정한 모델이었다. 연구대상은 정신분열병 환자 270명이었다. χ2 차이 검증 결과, 대안모델들에 비해 5개 요인간 상관을 가정한 제안모델이 적합한 것으로 나타났다. 또한 1주, 2주, 4주 간격 각각에서 Happy-QoL 척도에 대한 정신분열병 환자들의 반응은 시간적 안정성이 있었다. 끝으로, 본 연구의 시사점을 간략하게 논의하였다.

Happy-QoL Scale was developed to assess the quality of life in patients with schizophrenic disorder and composed 5 factors(economic domain, general domain, physical domain, occupational/social domain, and affective domain). This study was intended to confirm that the factor model of the Happy-QoL scale developed by explanatory factor analysis could be fitted to the newly collected data from the subjects and to examine the responses to the QoL scale had a time stability over the different time intervals. The principally proposed model was consisted of five factors which were correlated each other. The alternative models suggested were as follows: the model supposed no correlation among 5 factors; the model supposed correlation between factor 1(economic domain) and factor 4(occupational/social domain), and between factor 2(general domain) and factor 3(physical domain). The subjects were composed of 270 patients with schizophrenic disorder. The results of χ2 difference test showed that the model supposing five factors were correlated each other was best fitted. The responses to the Happy-QoL scale had the time stability on each periods of week 1, week 2, and week 4. Finally, the suggestions of this study were briefly described.

정확도가 높은 교통정보 예측은 지능형교통체계(intelligent transport systems, ITS)를 통한 교통 시설 이용자들의 혼잡 경로 회피 안내 등에서 활용되는 중요한 기능이다. 정확한 교통정보 예측을 위해 다양한 딥러닝 모델들이 발전되어 왔다. 최근에는 앙상블 기법을 활용하여 다양 한 모델들의 장단점을 결합하여 예측 정확도와 안정성을 높이고 있다. 따라서, 본 연구에서는 다양한 딥러닝 모델들을 활용하여 교통정보 예측 모델을 개발하였으며, 개발된 딥러닝 모델들 을 스태킹 앙상블(stacking ensemble)하여 성능을 개선하였다. 개별 모델들은 교통량 예측에서 10% 이내의 오차율을, 속도 예측에서 3% 이내의 오차율을 보였다. 앙상블 모델은 교차검증을 수행하지 않았을 때, 타 모델과 비교하여 더욱 높은 정확도를 보였다. 교차검증을 수행한 앙상 블 모델은 장기예측에서 타 모델보다 균일한 오차율을 보이는 것으로 나타났다.

Accurate traffic information prediction is considered to be one of the most important aspects of intelligent transport systems(ITS), as it can be used to guide users of transportation facilities to avoid congested routes. Various deep learning models have been developed for accurate traffic prediction. Recently, ensemble techniques have been utilized to combine the strengths and weaknesses of various models in various ways to improve prediction accuracy and stability. Therefore, in this study, we developed and evaluated a traffic information prediction model using various deep learning models, and evaluated the performance of the developed deep learning models as a stacking ensemble. The individual models showed error rates within 10% for traffic volume prediction and 3% for speed prediction. The ensemble model showed higher accuracy compared to other models when no cross-validation was performed, and when cross-validation was performed, it showed a uniform error rate in long-term forecasting.

본 연구에서는 고객생애가치와 고객 자산의 대표적 측정 방법 중 상대적으로 데이터 접근이 용이하여 많이 활용되고 있는 기업이 공개하는 각종 공시 자료를 활용한 간접적 측정 방법과 고객 설문을 통한 인지적 측정 방법을 동일 기업에 대해 적용하여 두 개의 이종 측정의 교차 타당성과 외적 타당성을 시장 가치를 활용하여 비교 분석하고 그 장단점과 두 측정 방법의 구조적 차이인 경쟁 요소의 고려 여부에 따른 오차 발생의 가능성도 동시에 탐색하였다. 고객 자산 측정 방법의 교차 비교를 위해 고객 자산 추정에 많이 활용되어온 통신 산업을 대상으로 자료 수집 및 분석을 수행하였다. 고객 자산의 측정과 일관된 비교를 위하여 먼저 이동통신 이용 패턴에 대해 전국(서울/수도권 및 6대 광역시)의 성인 남녀를 대상으로 지역별 인구 비례 할당을 통해 설문 조사가 진행되었다. 설문 조사 기간에 해당하는 기간 동안의 기업 공시 자료를 활용한 고객 자산이 추정되었다. 분석 결과, 두 측정 방법을 통한 고객 자산 측정값이 대표적 기업 가치로 알려진 시장 가치와 유사한 결과를 산출하고 있어 외적 타당성이 확인되었다. 특히, 고객 설문을 통한 측정 방법이 인지적 측정 방법의 단점에도 불구하고 오히려 시장 가치를 좀 더 정확하게 측정할 수 있는 가능성을 보여주었다. 또한, 시장의 경쟁 구조를 고려할 경우 각 측정 방법의 장단점이 다르게 표출됨을 알 수 있다. 즉, 기업이 공개하는 각종 공시 자료를 활용한 간접적 측정 방법의 경우 시장 내 경쟁에서 열위에 있는 기업의 경우 상대적으로 과대평가 되고 경쟁에서 우위에 있는 기업의 경우 고객 자산이 상대적으로 과소평가 되는 경향이 있었다. 하지만, 고객 설문을 통한 인지적 측정 방법의 이러한 과대 또는 과소평가 경향은 약화되고, 특히 열위에 있는 기업의 고객 자산 측정 결과는 시장 가치에 상당히 근접하는 측정 결과를 보여주었다. 전반적으로 분석 결과는 두 이종 측정 방법의 비교 결과 기업의 공시 자료를 활용한 고객생애가치와 고객 자산의 측정 방법보다 시장의 경쟁 상황을 적절히 반영할 수 있는 고객 설문 기반의 고객생애가치와 고객 자산의 측정 방법이 더 우월할 수 있는 가능성을 제시한다. 더불어, 시장 가치 대비 과소 또는 과대평가 경향을 고려할 때 고객생애가치와 고객 자산 측정에 있어서 시장 경쟁 요인은 상당히 중요한 측정 요인으로 작용할 수 있음을 제안한다.

The purpose of this study is to cross-verify the indirect measurement method using various publicly disclosed data and the cognitive measurement method through a customer survey. The cross validity and external validity of the two heterogeneous measurements were compared and analyzed using market values, and the possibility of error due to the consideration of competitive factors, which is the structural difference between the two measurement methods, was also explored. For cross-comparison of customer equity measurement methods, data collection and analysis were performed for the telecommunications industry, which has been widely used in customer equity estimation. For a consistent comparison with the measurement of customer equity, a survey was conducted on mobile communication usage patterns by region targeting all the customers across the country. Customer equity was estimated using corporate disclosure data during the period corresponding to the survey period. As a result of the analysis, the external validity of the customer's equity measurement through the two measurement methods was confirmed as the result was similar to the market value known as the representative corporate value. In particular, the cognitive measurement method through customer survey showed the possibility to measure the market value more accurately despite the disadvantages of the measurement method based on customer perception. Besides, it can be seen that the strengths and weaknesses of each measurement method are expressed differently when considering the competitive structure of the market. In other words, in the case of the indirect measurement method using various publicly disclosed data disclosed by companies, companies that are inferior in competition in the market tend to be relatively overestimated, and companies that are superior in competition tend to be relatively underestimated. However, this tendency to over or underestimate the cognitive measurement method through customer surveys is attenuated, and the measurement results of customer equity of particularly inferior companies have shown measurement results that are very close to the market value. Overall, the comparison result of the two heterogeneous measurement methods shows that the customer survey-based measurement method that can adequately reflect the market competition situation is better than the measurement method using publicly disclosed data of companies. Also, it is suggested that market competition factors can act as a very important measurement factor in measuring customer lifetime value and customer equity when considering the tendency of underestimating or overestimating the market value.

본 연구는 청소년 유대감 척도(Measure of Adolescent Connectedness)를 한국 청소년에 맞게 타당화하고자 하였다. 이를 위해 중고등학교 청소년 702명이 설문에 참여했고 타당화를 위해 한국판 청소년 유대감 척도, 학교적응 척도, 사회적 지지 척도, 자아존중감 척도가 사용되었다. 원 척도의 10요인 구조 모형이 맞는지 검증하기 위해 확인적 요인분석을 실시하였고 그 결과 10요인구조로 나타난 원 척도와 달리 한국판 유대감 척도는 7요인 구조가 더 적합한 것으로 나타났다. 원래 포함되어 있던 자신의 현재, 자신의 미래, 또래와의 유대요인을 제외한 7요인이-동네, 친구, 부모, 형제, 학교, 교사, 읽기와의 유대- 적합한 모형으로 지지되었다. 유대감 척도의 하위요인들의 내적 합치도는 양호하게 지지되었고 각 하위요인은 비슷한 척도들과 높은 정적 상관을 보였다. 청소년 남녀 집단을 비교한 결과, 남자 집단은 동네와의 유대감이 유의미하게 높았으며 여자집단은 학교, 교사, 부모와의 유대감이 유의미하게 높았다. 이러한 연구 결과들은 한국 청소년을 대상으로 유대감 척도가 유용한 도구로 사용될 수 있음을 시사해준다. 결과를 토대로 후속연구에 대한 제언과 연구의 제한점 등이 마지막으로 논의되었다.

The purpose of the study was to validate the Measure of Adolescent Connectedness(MAC: Karcher, 2005) to Korean adolescents. A total of 702 adolescents participated in the study. Based on the Yu's (2007) validation study, the original measure was translated to the Korean version. To examine the 10-factor construct of the original measure of MAC, a confirmatory factor analysis(CFA) was utilized. Results indicated that 7-factor model fits better responses to Korean version MAC than the original 10-factor model. The internal consistency coefficients for sub-factors within the Korean version MAC were at reasonable level. Convergent validity was examined through investigating correlations between the MAC and other scales (i.e., parent support, friend support, teacher support, and school adjustment). These results indicated that the 7-factor model of MAC is a reliable and valid measure for Korean adolescents. Limitations and suggestions for further study are discussed.