• Annual Conference on Human and Language Technology
• /
• 2014.10a
• /
• pp.112-116
• /
• 2014

서답형 문항은 학생들의 종합적인 사고능력을 판단하는데 매우 유용하지만 채점할 때, 시간과 비용이 매우 많이 소요되고 채점자의 공정성을 확보해야 하는 어려움이 있다. 이러한 문제를 개선하기 위해 본 논문에서는 서답형 문항에 대한 자동채점 시스템을 제안한다. 본 논문에서 제안하는 시스템은 크게 언어 처리 단계와 채점 단계로 나뉜다. 첫 번째로 언어 처리 단계에서는 형태소 분석과 같은 한국어 정보처리 시스템을 이용하여 학생들의 답안을 분석한다. 두 번째로 채점 단계를 진행하는데 이 단계는 아래와 같은 순서로 진행된다. 1) 첫 번째 단계에서 분석 결과가 완전히 일치하는 답안들을 하나의 유형으로 간주하여 각 유형에 속한 답안의 빈도수가 높은 순서대로 정렬하여 인간 채점자가 고빈도 학생 답안을 수동으로 채점한다. 2) 현재까지 채점된 결과와 모범답안을 학습말뭉치로 간주하여 자질 추출 및 자질 가중치 학습을 수행한다. 3) 2)의 학습 결과를 토대로 미채점 답안들을 군집화하여 분류한다. 4) 분류된 결과 중에서 신뢰성이 높은 채점 답안에 대해서 인간 채점자가 확인하고 학습말뭉치에 추가한다. 5) 이와 같은 방법으로 미채점 답안이 존재하지 않을 때까지 반복한다. 제안된 시스템을 평가하기 위해서 2013년 학업성취도 평가의 사회(중3) 및 국어(고2) 과목의 서답형 문항을 사용하였다. 각 과목에서 1000개의 학생 답안을 추출하여 채점시간과 정확률을 평가하였다. 채점시간을 전체적으로 약 80% 이상 줄일 수 있었고 채점 정확률은 사회 및 국어 과목에 대해 각각 98.7%와 97.2%로 나타났다. 앞으로 자동 채점 시스템의 성능을 개선하고 인간 채점자의 집중도를 높일 수 있도록 인터페이스를 개선한다면 국가수준의 대단위 평가에 충분히 활용할 수 있을 것으로 생각한다.

• The KIPS Transactions:PartB
• /
• v.14B no.3 s.113
• /
• pp.223-230
• /
• 2007

The purpose of developing an automated scoring system for English composition is to score the tests for writing English sentences and to give feedback on them without human's efforts. This paper presents an automated system to score English composition, whose input is a single sentence, not an essay. Dealing with a single sentence as an input has some advantages on comparing the input with the given answers by human teachers and giving detailed feedback to the test takers. The system has been developed and tested with the real test data collected through English tests given to the third grade students in junior high school. Two steps of the process are required to score a single sentence. The first process is analyzing the input sentence in order to detect possible errors, such as spelling errors, syntactic errors and so on. The second process is comparing the input sentence with the given answer to identify the differences as errors. The results produced by the system were then compared with those provided by human raters.

• Journal of The Korean Association For Science Education
• /
• v.43 no.3
• /
• pp.237-251
• /
• 2023

This study explores the possibility of automated scoring for scientific graph answers by designing an automated scoring model using convolutional neural networks and applying it to students' kinematics graph answers. The researchers prepared 2,200 answers, which were divided into 2,000 training data and 200 validation data. Additionally, 202 student answers were divided into 100 training data and 102 test data. First, in the process of designing an automated scoring model and validating its performance, the automated scoring model was optimized for graph image classification using the answer dataset prepared by the researchers. Next, the automated scoring model was trained using various types of training datasets, and it was used to score the student test dataset. The performance of the automated scoring model has been improved as the amount of training data increased in amount and diversity. Finally, compared to human scoring, the accuracy was 97.06%, the kappa coefficient was 0.957, and the weighted kappa coefficient was 0.968. On the other hand, in the case of answer types that were not included in the training data, the s coring was almos t identical among human s corers however, the automated scoring model performed inaccurately.

• Korean Journal of Culture and Social Issue
• /
• v.23 no.2
• /
• pp.295-332
• /
• 2017

The main purpose of this research is to analyze stress, social support and coping behavior of adults based on their level of self-efficacy. A total of 899 adults (399 male; 500 female), each with a child attending either elementary and secondary school, participated in the study. The inter-rater reliability for the open-ended questionnaire utilized in the study was 93.4%, with a Kappa coefficient of .92. The range of Cronbach α for the variables measured through a quantitative method was .87~.92. The results were as follows: First, the representative responses to the question about their most painful stress experiences were, financial difficulties, child rearing and duties of workplace. The Lower Efficacy group, compared to the Upper Efficacy group, responded much more with financial difficulties related responses. There were significant differences in the level of stress symptoms according to level of self-efficacy. The Lower Efficacy group expressed stronger levels of stress symptoms when compared to the Upper Efficacy group. Second, in terms of social support, the participants responded that they received the most help from their family members, followed by none(self), and friends. When comparing the two efficacy groups, the Upper Efficacy group responded most frequently that they received social support from their family members, whereas the Lower Efficacy group indicated none. There were significant differences in the level of relational conflicts according to the level of self-efficacy. The Upper Efficacy group showed much less conflict in parent-child relations, spousal relations and relations with their boss, compared to the Lower Efficacy group. Third, for the type of social support participants received, the most frequent response was emotional support, followed by none, and advice. Relatively, when comparing the two groups with each other, the Lower Efficacy group responded more frequently with none, whereas for the Upper Efficacy group responded more frequently with advice. There were significant differences in the amount of emotional support received according to level of self-efficacy. The Upper Efficacy group received much more emotional support from their spouses and their bosses compared to the Lower Efficacy group. Fourth, the most frequently adopted coping style to stress was self-regulation, followed by direct problem solving, and nothing(none). The most frequent response for the Upper Efficacy group was direct problem solving, whereas for the Lower Efficacy group was nothing(none). There was a significant difference in coping efficiency to stress according to level of self-efficacy. The Upper Efficacy group coped more efficiently with stress than the Lower Efficacy group.