• Journal of Korean Critical Care Nursing
• /
• v.15 no.3
• /
• pp.37-50
• /
• 2022

Purpose : This study aimed to examine the item characteristics of the Korean version of the intensive care experience questionnaire (K-ICEQ) using the Rasch analysis model of the item response theory. Methods : In this methodological study, the validity of the scale was examined, and a secondary analysis was conducted using cohort data of patients who were discharged from the intensive care units (ICU). Data from 891 patients who responded to the K-ICEQ upon ICU discharge were analyzed. The WINSTEP program was used to analyze item characteristics, including item difficulty, fit indices, appropriateness scale, and separation reliability. Results : The difficulty level of all 26 items of the K-ICEQ was appropriate, and the fit indices of the 25 items, except for item 18, were good. The 5-point scale of the K-ICEQ was not appropriate in the three subscales. The item separation reliability was good in all subscales, but did not meet the criteria in terms of respondents. Conclusion : The results of examining the item characteristics of the K-ICEQ revealed a good degree of difficulty, fitness, and item separation reliability. To increase the validity of the K-ICEQ, we suggest the rearrangement of the overall item order, modification of the item description of the "recall of experience" subscale, and reduction of the scale response level.

• Journal of Educational Research in Mathematics
• /
• v.14 no.1
• /
• pp.71-88
• /
• 2004

The purpose of this study was to examine particular variables that play a significant role in the difficulty of math test items in College Scholastic Ability Test (CSAT). The study also aimed to develop a model of measuring the item difficulty. Variables correlated to item difficulty were drawn from the review of the related literature and the analysis of the content and difficulty of the past test items of CSAT. The first instrument was designed by using the correlated variables. According to the results of correlation analysis, the second instrument was made by deleting the variables which showed relatively low correlation with item difficulty and by refining some variables. Several models were proposed by using the revised instrument. The comparison of the R square and cross validity of each model reveals that integrated regression model was the most stable and accurate among the proposed models. The study also showed that statistically significant predictors were choice format, content domain, behavior domain, and the degree of item familiarity in the order of proportion of variance accounted by the predictors. Despite the limited scope of the present research, it can be suggested that its findings provide useful insights into predicting math test item difficulty.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.249-263
• /
• 2023

Collaborative Filtering, a representative recommendation system methodology, consists of two approaches: neighbor methods and latent factor models. Among these, the latent factor model using matrix factorization decomposes the user-item interaction matrix into two lower-dimensional rectangular matrices, predicting the item's rating through the product of these matrices. Due to the factor vectors inferred from rating patterns capturing user and item characteristics, this method is superior in scalability, accuracy, and flexibility compared to neighbor-based methods. However, it has a fundamental drawback: the need to reflect the diversity of preferences of different individuals for items with no ratings. This limitation leads to repetitive and inaccurate recommendations. The Adaptive Deep Latent Factor Model (ADLFM) was developed to address this issue. This model adaptively learns the preferences for each item by using the item description, which provides a detailed summary and explanation of the item. ADLFM takes in item description as input, calculates latent vectors of the user and item, and presents a method that can reflect personal diversity using an attention score. However, due to the requirement of a dataset that includes item descriptions, the domain that can apply ADLFM is limited, resulting in generalization limitations. This study proposes a Generalized Adaptive Deep Latent Factor Recommendation Model, G-ADLFRM, to improve the limitations of ADLFM. Firstly, we use item ID, commonly used in recommendation systems, as input instead of the item description. Additionally, we apply improved deep learning model structures such as Self-Attention, Multi-head Attention, and Multi-Conv1D. We conducted experiments on various datasets with input and model structure changes. The results showed that when only the input was changed, MAE increased slightly compared to ADLFM due to accompanying information loss, resulting in decreased recommendation performance. However, the average learning speed per epoch significantly improved as the amount of information to be processed decreased. When both the input and the model structure were changed, the best-performing Multi-Conv1d structure showed similar performance to ADLFM, sufficiently counteracting the information loss caused by the input change. We conclude that G-ADLFRM is a new, lightweight, and generalizable model that maintains the performance of the existing ADLFM while enabling fast learning and inference.

• Journal of Korean Society for Quality Management
• /
• v.13 no.1
• /
• pp.56-64
• /
• 1985

This paper presents a general algorithm of multi-item continuous review models to obtain simultaneous solutions for ordering quantities and reorder points for each item in an inventory, while satisfying constraints on average inventory investment and reordering workload. Two models are formulated'in each model the heuristic method is utilized, and the partial back-logging is considered. In the first model, the objective function is the minimization of total inventory variable cost. In the second model, the objective function is the minimization of total time-weighted shortages, and the ordering, holding, and stockout costs in this model are independent each other. A numerical example is also solved to present application of each model.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.25 no.1
• /
• pp.42-48
• /
• 2002

This paper addresses the necessity of integration between buyer and suppliers for effective implementation of Just-In-Time purchasing in a multi-item environment. An integrated inventory model of facilitating multiple shipments in small lots is developed. Also, an iterative solution procedure is developed to simultaneously find the order(contract) interval for each item and number of shipments between buyer and suppliers. We show by example that when the integrated policy is adopted by both buyer and suppliers in a cooperative manner, both parties can benefit.

• International Journal of Reliability and Applications
• /
• v.17 no.1
• /
• pp.51-63
• /
• 2016

Stress function of ball bearing is function of multiple stochastic factors and this system is so complex that analytical expression for reliability is difficult to obtain. To address this pressing problem, in this article, we have made an attempt to approximate system reliability of this important item based on reliability bounds under the stress strength setup. This article also provides level of error of this item. Numerical analysis has been adopted to show the closeness between the upper and lower bounds of this item.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.15 no.26
• /
• pp.173-179
• /
• 1992

This paper is concerned with preventive maintenance model for the items whose failures are dependent on their wear level. The previous maintenance models have used time as their decision variable, but it is not appropriate for the case which have wear dependent processes for their failures. In this paper, we consider an operating item which is under periodic review and which is subject to degradation. The scheduled maintenance (overhaul ) is based on the level of item degradation rather time. A functional equation for the total expected cost over an infinite horizon period is formulated and solved.

• Korean Journal of Human Ecology
• /
• v.16 no.1
• /
• pp.159-171
• /
• 2007

The purpose of this research is to analyze the difference of Structural Equation Model which shows the path between BRQ and its performance variables according to purchase product types-fashion brand types, clothing item groups. The subjects were women in their 20s to 40s living in Seoul and Metropolitan areas, and 482 copies of questionnaire were analyzed. Multi-Group Analysis of AMOS 5.0 Package was used to investigate structural equation model according to fashion brand types and clothing item groups. The results of this study were as follows. As for fashion brand types, there appeared to be significant differences between high price brand type and medium-low price brand type for three paths, namely brand satisfaction to brand loyalty, BRQ to brand attitude, and brand attitude to brand loyalty. However the verification of structural equation model according to clothing item groups showed no significant differences between formal wear and informal wear. Consequently, BRQ was proved to affect brand satisfaction and brand loyalty, and brand satisfaction was the important intermediate variable between BRQ and brand loyalty. As consumers were likely to show the difference of structural equation model according to the price of purchase goods, differencial marketing strategy would be suggested.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.04b
• /
• pp.244-246
• /
• 2002

The rapid increase of E-commerce market requires a solution to assist the buyer to find his or her interested items. The intelligent agent model is one of the approaches to help the buyers in purchasing items in outline market. In this paper, the user query processing model in the item recommendation agent is proposed. In the proposed model, the retrieval result is affected by the automatically generated queries from user preference information as well as the queries explicitly given by user. Therefore, the proposed model can provide the customized search results to each user.

• Korean Security Journal
• /
• no.42
• /
• pp.37-57
• /
• 2015

The purpose of this study was to verify goodness-of-fit by Many-Facets Rasch model for applying martial arts training satisfaction scale to security martial arts trainees. To achieve the purpose, 255 security martial arts trainees' data were used in the analysis. In addition, In addition, the AMOS 20.0 program was used for unidimensionality validation, and take advantage of the Facets 3.61 program for goodness-of-fit verification. Specific results are as follows: First, Unidimensionali test results showed that model fit, reliability and standardized ${\beta}$ value are suitable. Second, the analysis results of goodness-of-fit, items 1, 2, 3 are inadequate, 4, 8, 11, 13, 14 items once found to be over-fit questions. Also, analysis of item difficulty, item 1 has highest difficulty and item 7 was lowest. Third, According Facets item difficulty and response difference verification result, female group exhibited a high level of item difficulty compared to the male group, goodness-of-fit was all accurate. As the result of item difficulty and response difference verification based on Martial arts training flow, there is no response difference according to the training experience. On the other hands, less than 4 years to 5 years and less than 5 years to 6 years trainees represented over-fit features. Results of item difficulty and response difference verification by grade level, first grade was the most highly recognized the item difficulty and fourth grade was also recognized the lowest of item difficulty Fourth, the response category analysis showed that the six points response categories are not appreciate.