• 한국농림기상학회지
• /
• 제23권4호
• /
• pp.329-339
• /
• 2021

콩은 논 대표적인 밭작물로써 온도, 수분, 토양과 같은 환경 조건에 민감하기 때문에 재배 시 포장 관리가 매우 중요하다. 작물 상태를 비파괴적, 비접촉적 방법으로 측정할 수 있는 분광 기술을 활용한다면 작황 예측, 작물 스트레스 및 병충해 판별 등 생육 진단 및 처방을 통해 품질과 수확량을 높일 수 있다. 본 연구에서는 회전익 무인기에 탑재된 다중분광 센서를 이용하여 시험 포장에서 콩 생육 추정 모델 개발하고 재현성을 확인하기 위해 농가 포장에 검증을 수행하였다. 분광 데이터로 산출된 정규화 식생지수(NDVI, GNDVI), 단순비 식생지수(RRVI, GRVI)와 콩 생육 데이터(생체중, LAI)를 선형회귀분석을 실시하여 모델을 개발하였으며 괴산에 위치한 농가포장에서 검증을 실시하였다. 그 결과 생체중의 경우 정규화 식생지수를 이용 시 포화되기 때문에 단순비 식생지수 GRVI를 이용한 모델의 성능이 가장 높았다(R²=0.74, RMSE=246 g/m², RE=34.2%). 괴산 농가 포장에 생체중 모델 검증 결과 RMSE=392 g/m², RE=32%로 나타났으며 작부 체계별 나누어 검증 결과 단작 포장과 이모작 포장 생체중 모델은 RMSE=315 g/m², RE=26% 및 RMSE=381 g/m², RE=31%로 나타났다. 작부 체계별 포장과 적산온도가 유사한 연도별 시험 포장(2018+2020년, 2019년)을 나누어 생체중 모델 개발한 결과 단년도(2019년)의 성능이 높게 나타났다. 작부 체계별 적산온도가 유사한 검증과 기존 검증 간 비교 결과 단작 포장은 RMSE 및 RE를 기준으로 각각 29.1%와 34.3%로 개선되었으나 이모작 포장은 -19.6%, -31.3%로 저하되었다. 적산온도 이외의 환경 요인, 분광 및 생육 데이터 추가 시 다양한 환경 조건에서 재배되는 콩 생육을 추정 가능할 것으로 판단된다.

• 지능정보연구
• /
• 제22권3호
• /
• pp.23-43
• /
• 2016

최근 IT기술의 발전에 따라 많은 사람들이 자신들의 여가활동에 대한 경험을 공유하고 있으며, 역으로 다른 사람들의 여가활동에 대한 경험을 참고하여 더 나은 여가활동을 누릴 수 있는 기회를 얻게 되었다. 이러한 현상은 영화, 숙박, 음식, 여행 등 여가활동 전반에 걸쳐 나타나고 있으며, 그 중심에는 여가활동에 대한 정보를 요약하여 제공하는 수많은 사이트가 있다. 대부분의 여가활동 정보 사이트는 각 상품에 대한 평균 평점뿐만 아니라 상세 리뷰를 제공함으로써, 해당 상품을 구매하고자 하는 잠재고객의 의사결정을 지원하고 있다. 하지만 기존 대부분의 사이트는 한 단계의 평가기준에 따라 평점과 리뷰를 제공하기 때문에, 각 평가기준을 구성하는 세부요소에 대한 특징과 평가기준 별 주요 이슈를 파악하기 위해서는 상당히 많은 수의 리뷰를 직접 읽어야 한다는 불편이 따른다. 즉 사용자는 자신이 중요한 것으로 생각하는 평가기준에 대한 조건을 파악하기 위해, 많은 수의 리뷰를 하나하나 읽어보는 과정에서 많은 시간과 노력을 소비하게 된다. 예를 들어 호텔의 접근성, 객실, 서비스, 음식 등 한 단계의 평가기준만을 사용하여 평점과 리뷰를 제공하는 사이트의 경우, 접근성 중 특히 지하철역과의 거리, 객실 중 특히 욕실의 상태를 살펴보고자 하는 사용자에게 필요한 정보를 충분히 제공하지 못하게 된다. 따라서 본 연구에서는 기존 여가활동 정보 사이트의 한계, 즉 평가기준별로 입력된 리뷰를 신뢰하기 어렵다는 점과 평가기준을 구성하고 있는 세부 내용을 파악하기 어렵다는 점을 극복하기 위한 방안을 제시하고자 한다. 본 연구에서 제안하는 방법론은 사용자가 별도의 구분 없이 입력한 리뷰를 그 내용에 따라 평가기준별로 자동 분류하고, 각 평가 기준 별 주요 이슈를 요약하여 제공한다. 제안 방법론은 최근 텍스트 분석에 활발하게 사용되고 있는 토픽 모델링(Topic Modeling)에 기반을 두고 있으며, 각 리뷰를 하나의 문서 단위로 사용하는 것이 아니라 리뷰를 문장 단위로 끊어 개별 리뷰 유닛(Review Unit)으로 분해한 뒤, 평가기준별로 리뷰 유닛을 재구성하여 분석한다는 측면에서 기존의 토픽 모델링 기반 연구와 큰 차이가 있다고 할 수 있다. 본 논문에서는 제안 방법론을 실제 호텔 정보 사이트에서 수집한 423건의 리뷰 문서에 적용하여 6가지 평가기준에 대해 총 4,860건의 리뷰 유닛을 재구성하고, 이에 대한 분석 결과를 소개함으로써 제안 방법론의 유용성을 간접적으로 보인다.

• 지능정보연구
• /
• 제18권4호
• /
• pp.59-77
• /
• 2012

기업의 성장성, 수익성, 안정성, 활동성, 생산성 등에 대한 다양한 분석이 은행, 신용평가기관, 투자자 등 많은 이해관계자에 의해 실시되고 있고, 이에 대한 다양한 경영분석 지표들 또한 정기적으로 발표되고 있다. 본 연구에서는 이러한 경영분석 지표를 이용하여 어떤 기업이 가까운 미래에 유상증자를 실시하는지를 데이터마이닝을 통해 예측하고자 한다. 본 연구를 통해 어떠한 지표가 유상증자 여부를 예측하는데 도움이 되는가를 살펴 볼 것이며, 그 지표들을 이용하여 예측할 경우 그 예측의 정확도가 어느 정도인지를 분석하고자 한다. 특히 1997년 IMF 금융위기 전후로 유상증자를 결정하는 변수들이 변화하는지, 그리고 예측의 정확성에 분명한 차이가 존재하는지 분석한다. 또한 유상증자 실시 시기를 경영분석 지표 발표 후 1년 내, 1~2년 내, 2~3년 내로 나누어 예측 시기에 따라 예측의 정확성과 결정 변수들의 차이가 존재하는지도 분석한다. 658개의 유가증권상장법인의 경영분석 데이터를 이용하여 실증 분석한 결과, IMF 이후의 유상증자 예측모형이 IMF 이전의 예측모형에 비해 예측 정확도가 높았고, 학습용 데이터의 예측 정확도와 검증용 데이터의 예측 정확도 차이도 IMF 이후가 낮게 나타났다. 이러한 결과는 IMF 이후 재무자료의 정확도가 높아졌고, 기업에게 유상증자의 목적이 더욱 명확해졌다고 해석될 수 있다. 또한 예측기간이 단기인 경우 경영분석 지표 중 안전성에 관련된 지표들의 중요성이 부각되었고, 장기인 경우에는 수익성과 안전성뿐만 아니라 활동성과 생산성 관련지표도 유상증자를 예측하는 데 중요한 것으로 파악되었다. 그리고 모든 예측모형에서 산업코드가 유상증자를 예측하는 중요변수로 포함되었는데 이는 산업별로 서로 다른 유상증자 유형이 존재한다는 점을 시사한다. 본 연구는 투자자나 재무담당자가 유상증자 여부를 장단기 시점에서 예측하고자 할 때 어떠한 경영분석지표를 고려하여 분석하는 것이 바람직한지에 대한 지침을 제공하는데 그 의의가 있다.

• 핵의학기술
• /
• 제15권2호
• /
• pp.111-115
• /
• 2011

체외 검사는 각 검사 마다 6-7개의 standard 농도를 이용한 표준곡선(Dose response curve)을 사용하여 검사 결과를 분석하며, 표준곡선 사용은 진료재료 소모에 많은 부분을 차지하고 있다. 현재 원내에서 검사하는 검사항목 가운데 일부 검사 항목은, standard 농도를 이용한 환자의 검사 결과 분석수치가 대부분 낮은 결과의 분포를 보이고 있었으며, 이러한 검사 들은 마지막 standard 농도의 사용이 불필요하다고 판단하였다. 이에 낮은 검사 결과의 분포를 보이는 검사항목들을 선정하고, 마지막 standard 농도의 자리에 사용했던 진료 재료를 절감하여 진료재료의 소모비용을 줄였다. 원내 검사 항목 중 11개의 검사 항목을 진료재료 절감 대상항목으로 하였고, 검사가 진행한 횟수 만큼 마지막 standard농도에 사용되었던 진료재료(검사 Tube)를 절감할 수 있었다. 2009년 7월부터 2011년 2월까지 월별 진료재료 절감 개수, 절감 개수를 진료재료 단가로 계산하여 진료재료 절감액과, 절감된 진료재료를 환자 검사용 진료재료로 사용하여 발생한 수익을 ABC 원가계산을 근거로 산정하였다. 2009년 7월부터 2011년 2월까지 절감된 진료 재료는 3,131개였다. 이는 약 31kit의 검사 시약에 해당되는 진료재료를 절감하였다. 이것을 진료재료 단가로 계산하면 약 640만원의 진료재료비가 절감되었고, 이러한 진료 재료는 환자 검사 진료재료로 사용되어 약 1,370만원의 수익이 발생하였다. 환자의 검사 결과가 정상이거나, 낮게 분포하는 검사항목들은 마지막 standard 농도를 사용하지 않고 범위를 줄여서 검사를 진행하여도 검사 결과를 분석하는데는 영향이 없었고, standard 농도 범위 조정이 검사 결과의 신뢰성에는 영향이 없는 것으로 분석 되었다. 따라서 이 활동은 핵의학 체외검사의 진료재료 사용에 효율성을 높이고 진료재료 절감 및 병원 수익에 많은 효과가 있을 것으로 생각한다.

• 대한간호학회지
• /
• 제3권3호
• /
• pp.1-14
• /
• 1973

This study is designed to meet the following objectives: (1) To study attitude and behavior regarding marriage and age at marriage, (2) To learn correlates of age at marriage and to examine their relations, (3) To measure relative importance of the correlates of age at marriage, and (4) To study relations of age at marriage and family planning practice to fertility and their relative importance as correlates of fertility. The data are obtained by an independent cross-sectional survey in three study areas purposively selected to represent metropolitan. semihuman. rural population. The study population is confined to women age 17-50 as of survey. The overall response rate is 90%. Reliability of data is measured by . individual and aggregate inconsistency based upon a 15% subsample of the original interviews. The individual inconsistency (31%) is found to be high compared to the aggregate inconsistency (6%) for all 85 variables. However, the magnitude of differences between means is small, and the mean absolute shifts and proportional shifts are also small on the whole. In a word respondents did not change their answers too extremely or radically. The study populations of each study area are compared on some basic characteristics. It is found that the three study populations have more dissimilarities than similarities. The findings on seven different attitudinal positions of women toward marriage indicate that there have been tremendous changes in all study areas Iron "traditional" attitudes which have been prevalent for a long time in Korean society to "liberalized" or "modernized" attitudes. An apparent tendency is that women generally take a position of a "golden mean" attitude by not preferring either extreme of marriage attitudes. Nevertheless, the young, single, educated, and urbanite appears more "liberalized. " There has been some increase in ideal age at marriage from 1958 to 1970 for both sexes. No age group, marital status, or study area differentials in ideal age at marriage are found, the average ideal age at marriage in every sub-group being 24-25. Awareness of existing legal marriageable ages is low; only 4.4% are aware that "with parental permission: minimum age for males is 18 years and for females 16 years,"and only 3.7% are aware that "without parental permission: 27 years for males and 23 years for females." People in Korra tend to marry spouses who are in various social ways like themselves: the similarities include (a) education, occupational status of father, (c) economic status, (d) usual residence before marriage, and (e) religion. Both singulars and actual mean ages at marriage in this study confirm the trend of rising age at marriage previously established by other independent studies. The urban-rural differential in age at marriage is observed, but the differential narrows down gradually from 1935 to 1970. All socio-economic, demographic, and other variables pertaining to wife before and at first marriage, excluding (a) religion, (b) father′s of occupation, and (c) as: of menarche, are correlated with respondent's age at first marriage, whereas only three variables out of all socio-economic variables relating to husband before and at wife′s first marriage, viz., (a) education, (b) usual residence, and (c) economic level of his old home, are correlated with respondent′s age at marriage. Among socio-economic and modernity variables related to either husband or wife at the time of survey, only education and duration of residence are correlated with wife′s age at first marriage. Among the correlates of respondent′age at first marriage, education is in general the most important variable. However, it is found that wife′s education is more important than husband′s. The combined effects or the correlates studied explain no more than about 40% of variance for any of the selected groups of variables. Points which might counteract the effects of late marriage on fertility are not serious in Korea. For each of the correlates of the three fertility indices chosen for this study. namely, (a) number of living children, (b) number of live births, and (c) number of pregnancies, age at marriage is the major contributor to the variance in all age groups except the age group of 20-29 in which the index of family planning practice is the major contributor. The proportion of variability in fertility indices accounted for by the correlates is never more than 40% of the total variance in any age group. Based upon the findings from this study, it could be concluded that in the foreseeable future (a) celibate group will no! be increased to a point that would slow down population growth rate in Korea, (b) age at marriage will not increase continually, (c) although education stands out as the major contributing variable which independently explains the variation in age at marriage, it seems probable that education may not be the major variable in the near future, and (d) despite the fact found by this study that age at marriages has been the major contributor to the variance of each of the fertility indices used, family planning practice will play a more important role in the reduction of fertility in the Korean society. Therefore, factors interrupting practice of family planning must be eliminated and family planning program should be strengthened if further fertility reduction is needed.

• 간호행정학회지
• /
• 제2권1호
• /
• pp.97-114
• /
• 1996

INSTRODUCTION Service quality is, unlike goods quality, an abstract and elusive constuct. Service quality and its requirements are not easily understood by consumers, and also present some critical research problems. However, quality is very important to marketers and consumers in that it has many strategic benefits in contributing to profitability of marketing activities and consumers' problem-solving activities. Moreover, despite the phenomenal growth of medical service sector, few researchers have attempted to define and model medical service quality. Especially, little research has focused on the evaluation of medical service quality and patient satisfaction from the perspectives of both the provider and the patient. As competition intensifies and patients are demanding higher quality of medical service, medical service quality and patient satisfaction has emerged as a critical research topic. The major purpose of this article is to explore the concept of medical service quality and its evaluation from both nurse and patient perspectives. This article attempts to achieve its purpose by (1)classfying critical service attibutes into threecategories(satisfiers, hygiene factors, and performance factors). (2)measuring the relative importance of need criteria, (3)evaluating SERVPERF model and SERVQUAL model in medical service sector, and (4)identifying the relationship between perceived quality and overall patient satisfaction. METHOD Data were gathered from a sample of 217 patients and 179 nurses in Seoul-area general hospitals. From the review of previous literature, 50 survey items representing various facets of the medical service quality were developed to form a questionnaire. A five-point scale ranging from "Strongly Agree"(5) to "Strongly Disagree"(1) accompanied each statement(expectation statements, perception statements, and importance statements). To measure overall satisfaction, a seven-point scale was used, ranging from "Very Satisfied"(7) to "Very Dissatisfied"(1) with no verbal labels for scale points 2 through 6 RESULTS In explaining the relationship between perceived performance and overall satisfaction, only 31 variables out of original 50 survey items were proven to be statistically significant. Hence, a penalty-reward analysis was performed on theses 31 critical attributes to find out 17 satisfiers, 8 hygiene factors, and 4 performance factors in patient perspective. The role(category) of each service quality attribute in relation to patient satisfaction was com pared across two groups, that is, patients and nurses. They were little overlapped, suggesting that two groups had different sets of 'perceived quality' attributes. Principal components factor analyses of the patients' and nurses' responses were performed to identify the underlying dimensions for the set of performance(experience) statements. 28 variables were analyzed by using a varimax rotation after deleting three obscure variables. The number of factors to be extracted was determined by evaluating the eigenvalue scores. Six factors wereextracted, accounting for 57.1% of the total variance. Reliability analysis was performed to refine the factors further. Using coefficient alpha, scores of .84 to .65 were obtained. Individual-item analysis indicated that all statements in each of the factors should remain. On 26 attributes of 31 critical service quality attributes, there were gaps between actual patient's importance of need criteria and nurse perceptions of them. Those critical attributes could be classified into four categories based on the relative importance of need criteria and perceived performance from the perspective of patient. This analysis is useful in developing strategic plans for performance improvement. (1) top priorities(high importance and low performance) (in this study)- more health-related information -accuracy in billing - quality of food - appointments at my convenience - information about tests and treatments - prompt service of business office -adequacy of accommodations(elevators, etc) (2) current strengths(high importance and high performance) (3)unnecessary strengths(low importance and high performance) (4) low priorities(low importance and low performance) While 26 service quality attributes of SERPERF model were significantly related to patient satisfation, only 13 attributes of SERVQUAL model were significantly related. This result suggested that only experience-based norms(SERVPERF model) were more appropriate than expectations to serve as a benchmark against which service experiences were compared(SERVQUAL model). However, it must be noted that the degree of association to overall satisfaction was not consistent. There were some gaps between nurse percetions and patient perception of medical service performance. From the patient's viewpoint, "personal likability", "technical skill/trust", and "cares about me" were most significant positioning factors that contributed patient satisfaction. DISCUSSION This study shows that there are inconsistencies between nurse perceptions and patient perceptions of medical service attributes. Also, for service quality improvement, it is most important for nurses to understand what satisfiers, hygiene factors, and performance factors are through two-way communications. Patient satisfaction should be measured, and problems identified should be resolved for survival in intense competitive market conditions. Hence, patient satisfaction monitoring is now becoming a standard marketing tool for healthcare providers and its role is expected to increase.

• Journal of Animal Science and Technology
• /
• 제45권2호
• /
• pp.183-190
• /
• 2003

한우의 고급육과 관련된 비육생리를 규명하여 최적의 출하시기를 결정하고 한우 개량을 위한 종축검정에 참고자료를 제공하기 위하여1999-2000년도에 출생하고 한우 후대검정에 공시된 거세 검정우들 중 127두를 무작위 추출하여 생후 11개월령부터 24개월령 출하시까지 1개월 단위로 생체단층촬영기를 이용하여 육질을 조사하였다. 조사된 자료를 이용하여 월령별 또는 체중 증가에 따른 배장근단면적, 등지방두께 및 근내지방도의 변화를 살펴보았고 이들 측정치가 도축하여 조사된 실측치와 상관관계를 알아 보았다. 분석결과 배장근단면적의 증가는 체중의 증가와 상당히 밀접하게 관계되었으며(P<0.01) 등지방두께의 증가는 비육후기로 갈수록 월령의 이차항과 정의 관계를 보였다(P<0.01). 근내지방도의 증가는 개체의 나이와 밀접한 관계를 보였으며(P<0.01) 개체간의 변이는 생후 11개월령부터 시작되고 있었으나 17개월령 이후로 갈수록 개체간 변이 폭이 커지고 있음을 알 수 있었다. 월령별 육질 측정치가 실측정치와의 상관관계에 있어서는 나이 및 체중에 대한 보정을 실시하지 않았을 경우, 생후 17개월령 이후에 조사된 전형질에 있어서 0.6이상의 상관관계를 보이고 있었으며 22개월령 이후의 측정치와 실측치 간에는 90% 전후의 높은 상관관계를 보이고 있었다. 출하 2주 이내의 초음파 측정치와 도축 실측치와의 상관은 나이 및 체중에 대한 보정을 실시하고 추정하였을 때, 배장근단면적에 대하여는 0.91~9.92, 등지방두께에서 0.94 및 근내지방도에서 0.88으로 비교적 높게 추정되었다. 하지만 분석에 고려된 육질 관련 월령별 초음파 측정치들간에 상관은 비교적 낮게 추정되었는데 이는 적은 조사자료 및 초음파 측정의 오차 등에 기인한 것으로 사료되며 보다 많은 연구가 필요할 것으로 사료되었다. 또한 한우에 있어서 생축으로부터 육질을 판정하기 위해서는 15-17개월령 이후가 적당할 것으로 사료되었으며 22개월령 전후의 측정치는 출하시기를 결정하는데 좋은 참고자료로 사용될 수 있을 것으로 사료되었다. 한편 종축선발을 위한 육질 조기판정은 외국의 12개월령 측정치와는 달리 15개월령 측정치가 보다 타당하다고 사료되었으나 이들 측정치에 대한 유전적 변이에 대한 연구가 필요할 것으로 사료되었다.

• 한국재료학회:학술대회논문집
• /
• 한국재료학회 2011년도 춘계학술발표대회
• /
• pp.14-14
• /
• 2011

W (tungsten)-alloys will be the most promising plasma facing armor materials in highly loaded plasma interactive components of the next step fusion reactors due to its high melting point, high sputtering resistance and low deuterium/tritium retention. The bonding technology of tungsten to Cu alloy was one of the key issues. In this paper, W/CuCrZr diffusion bonding has been performed successfully by inserting pure metal interlay. The joint microstructure, interfacial elements migration and phase composition were analyzed by SEM, EDS, XRD, and the joint shear strength and micro-hardness were investigated. The mock-ups were fabricated successfully with diffusion bonding and the cladding technology respectively, and the high heat flux test and thermal fatigue test were carried out under actively cooling condition. When Ni foil was used for the bonding of tungsten to CuCrZr, two reaction layers, Ni4W and Ni(W) layer, appeared between the tungsten and Ni interlayer with the optimized condition. Even though Ni4W is hard and brittle, and the strength of the joint was oppositely increased (217 MPa) due primarily to extremely small thicknesses (2~3 ${\mu}m$). When Ti foil was selected as the interlayer, the Ti foil diffused quickly with Cu and was transformed into liquid phase at $1,000^{\circ}C$. Almost all of the liquid was extruded out of the interface zone under bonding pressure, and an extremely thin residual layer (1~2 ${\mu}m$) of the liquid phase was retained between the tungsten and CuCrZr, which shear strength exceeded 160 MPa. When Ni/Ti/Ni multiple interlayers were used for bonding of tungsten to CuCrZr, a large number of intermetallic compound ($Ni_4W/NiTi_2/NiTi/Ni_3T$) were formed for the interdiffusion among W, Ni and Ti. Therefore, the shear strength of the joint was low and just about 85 MPa. The residual stresses in the clad samples with flat, arc, rectangle and trapezoid interface were estimated by Finite Element Analysis. The simulation results show that the flat clad sample was subjected maximum residual stress at the edge of the interface, which could be cracked at the edge and propagated along the interface. As for the rectangle and trapezoid interface, the residual stresses of the interface were lower than that of the flat interface, and the interface of the arc clad sample have lowest residual stress and all of the residual stress with arc interface were divided into different grooved zones, so the probabilities of cracking and propagation were lower than other interfaces. The residual stresses of the mock-ups under high heat flux of 10 $MW/m^2$ were estimated by Finite Element Analysis. The tungsten of the flat interfaces was subjected to tensile stresses (positive $S_x$), and the CuCrZr was subjected to compressive stresses (negative $S_x$). If the interface have a little microcrack, the tungsten of joint was more liable to propagate than the CuCrZr due to the brittle of the tungsten. However, when the flat interface was substituted by arc interfaces, the periodical residual stresses in the joining region were either released or formed a stress field prohibiting the growth or nucleation of the interfacial cracks. Thermal fatigue tests were performed on the mock-ups of flat and arc interface under the heat flux of 10 $MW/m^2$ with the cooling water velocity of 10 m/s. After thermal cycle experiments, a large number of microcracks appeared at the tungsten substrate due to large radial tensile stress on the flat mock-up. The defects would largely affect the heat transfer capability and the structure reliability of the mock-up. As for the arc mock-up, even though some microcracks were found at the interface of the regions, all microcracks with arc interface were divided into different arc-grooved zones, so the propagation of microcracks is difficult.

• 벤처창업연구
• /
• 제15권5호
• /
• pp.13-35
• /
• 2020

스타트업 액셀러레이터는 초기창업자들이 겪는 애로사항을 직접적으로 해결해 줄 수 있는 교육, 멘토링, 네트워킹, 공간, 시드머니(Seed money)등을 제공하고 일정 지분을 받는 새로운 형태의 투자주체이다(Clarysse, 2016). 스타트업 액셀러레이터들은 이들이 스타트업 생태계에 미치는 영향력이 입증되면서 전 세계적으로 확산되고 있다(Pauwels et al., 2016; Cohen & Hochberg, 2014). 본 연구는 이처럼 전 세계적으로 주요한 투자주체로 부상하고 있는 스타트업 액셀러레이터들의 투자결정요인을 도출하는 연구를 진행하였다. 먼저, 관련 선행연구가 부재한 상황이므로 정성적 메타합성법, 관련 자료조사, 관찰, 심층인터뷰의 과정을 거쳤다. 우선 창업초기 기업에 대한 투자결정요인에 대한 선행연구 30여 편을 선정하여 해당연구에 나타난 투자결정요인들을 정성적 메타합성법을 통해 비교·분석하였다. 그리고 액셀러레이터의 특성을 반영하기 위해서 액셀러레이터들의 평가과정에 대한 자료들을 조사하였으며, 실제로 이루어진 미국 액셀러레이터들의 스타트업 평가과정을 관찰하였다. 이 때 나타난 100여개의 질문을 근거이론의 코딩 기법을 활용하여 분석하였다. 더불어 미국 액셀러레이터들을 대상으로 심층 인터뷰를 진행하여 액셀러레이터의 특성과 주요 투자결정요인 등을 파악하였다. 이를 통해 최종적으로 5개 카테고리, 26개 세부 투자결정요인으로 이루어진 '액셀러레이터 형 투자결정요인'을 도출하였다. 그리고 이의 결과를 국내 액셀러레이터들의 검토를 받고 확정한 후, 국내 액셀러레이터를 대상으로 한 정량조사를 통해 각 카테고리 내, 중요도 및 우선순위를 살펴봄으로써 이의 신뢰성 및 액셀러레이터들이 중요하게 생각하는 투자결정요인 등을 확인하였다. 이렇게 액셀러레이터의 투자결정요인을 체계적으로 도출한 연구는 최초의 연구로써 후속연구들의 진행 및 스타트업 선정과정, 액셀러레이터의 투자결정 과정 등에 주요한 기여를 할 수 있으리라 기대한다.

• Asia pacific journal of information systems
• /
• 제21권4호
• /
• pp.1-25
• /
• 2011

Internet-based commerce has undergone an explosive growth over the past decade as consumers today find it more economical as well as more convenient to shop online. Nevertheless, the shift in the common mode of shopping from offline to online commerce has caused consumers to have worries over such issues as private information leakage, online fraud, discrepancy in product quality and grade, unsuccessful delivery, and so forth, Numerous studies have been undertaken to examine the role of perceived risk as a chief barrier to online purchases and to understand the theoretical relationships among perceived risk, trust and purchase intentions, However, most studies focus on empirically investigating the effects of trust on perceived risk, with little attention devoted to the effects of perceived risk on trust, While the influence trust has on perceived risk is worth studying, the influence in the opposite direction is equally important, enabling insights into the potential of perceived risk as a prohibitor of trust, According to Pavlou (2003), the primary source of the perceived risk is either the technological uncertainty of the Internet environment or the behavioral uncertainty of the transaction partner. Due to such types of uncertainty, an increase in the worries over the perceived risk may negatively affect trust, For example, if a consumer who sends sensitive transaction data over Internet is concerned that his or her private information may leak out because of the lack of security, trust may decrease (Olivero and Lunt, 2004), By the same token, if the consumer feels that the online merchant has the potential to profit by behaving in an opportunistic manner taking advantage of the remote, impersonal nature of online commerce, then it is unlikely that the merchant will be trusted, That is, the more the probable danger is likely to occur, the less trust and the greater need to control the transaction (Olivero and Lunt, 2004), In summary, a review of the related studies indicates that while some researchers looked at the influence of overall perceived risk on trust level, not much attention has been given to the effects of different types of perceived risk, In this context the present research aims at addressing the need to study how trust is affected by different types of perceived risk, We classified perceived risk into six different types based on the literature, and empirically analyzed the impact of each type of perceived risk upon consumer trust in an online merchant and further its impact upon purchase intentions. To meet our research objectives, we developed a conceptual model depicting the nomological structure of the relationships among our research variables, and also formulated a total of seven hypotheses. The model and hypotheses were tested using an empirical analysis based on a questionnaire survey of 206 college students. The reliability was evaluated via Cronbach's alphas, the minimum of which was found to be 0.73, and therefore the questionnaire items are all deemed reliable. In addition, the results of confirmatory factor analysis (CFA) designed to check the validity of the measurement model indicate that the convergent, discriminate, and nomological validities of the model are all acceptable. The structural equation modeling analysis to test the hypotheses yielded the following results. Of the first six hypotheses (H1-1 through H1-6) designed to examine the relationships between each risk type and trust, three hypotheses including H1-1 (performance risk ${\rightarrow}$ trust), H1-2 (psychological risk ${\rightarrow}$ trust) and H1-5 (online payment risk ${\rightarrow}$ trust) were supported with path coefficients of -0.30, -0.27 and -0.16 respectively. Finally, H2 (trust ${\rightarrow}$ purchase intentions) was supported with relatively high path coefficients of 0.73. Results of the empirical study offer the following findings and implications. First. it was found that it was performance risk, psychological risk and online payment risk that have a statistically significant influence upon consumer trust in an online merchant. It implies that a consumer may find an online merchant untrustworthy if either the product quality or the product grade does not match his or her expectations. For that reason, online merchants including digital storefronts and e-marketplaces are suggested to pursue a strategy focusing on identifying the target customers and offering products that they feel best meet performance and psychological needs of those customers. Thus, they should do their best to make it widely known that their products are of as good quality and grade as those purchased from offline department stores. In addition, it may be inferred that today's online consumers remain concerned about the security of the online commerce environment due to the repeated occurrences of hacking or private information leakage. Online merchants should take steps to remove potential vulnerabilities and provide online notices to emphasize that their website is secure. Second, consumer's overall trust was found to have a statistically significant influence on purchase intentions. This finding, which is consistent with the results of numerous prior studies, suggests that increased sales will become a reality only with enhanced consumer trust.