• Journal of Korean Society of Transportation
• /
• v.11 no.2
• /
• pp.5-26
• /
• 1993

In order to analyze travel mode choice behavior, behavioral models including logit model, based on revealed preference theory, have been using easily measurable variables such as individual socioeconomic characteristics and physical attributes of travel modes. But some recent attitudinal models of travel choice behavior have implied that the negligence of individual psychological variables and individual choice constraints in travel mode choice might preclude better prediction of individual travel mode choice behavior. In this context, this study was attempted to reconstruct an attitudinal model(AM), especially focused on the decision rules in travel mode choice decision making process, consistent with the conceptual framework relating individual attitude and choice constraints to choice behavior. And to evaluate the strengths of the AM to other comparative models(logit, linear-additive, conjunctive, lexicographic model) in predicting travel mode choice bebavior, an empirical study of the mode choice in work-trip to CBD in Seoul was performed. According to the results the percent of correct prediction(PCP) derived from the AM was higher than those derived from comparative models by at least 7 to 20% in predicting travel mode choice. But each model produced a different prediction accuracy depending on market segmentation by travel modal users, individual socioeconomic characteristics, transportation system characteristics, and satisfaction levels. The finding that different groups divided by a certain criterion employ different decision rules supports the necessity of developing a choice model such as the AM combining compensatory and noncompensatory decision rules, and suggests that a proposed transportation system management plan or policy may have different effects on each group.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.34 no.3
• /
• pp.29-41
• /
• 2009

The purpose of this paper is to develop a new classification method for predicting consumer choice based on genetic algorithm, and to validate Its prediction power over existing methods. To serve this purpose, we propose a hybrid model, and discuss Its methodological characteristics in comparison with other existing classification methods. Also, we conduct a series of experiments employing survey data of consumer choices of MP3 players to assess the prediction power of the model. The results show that the suggested model in this paper is statistically superior to the existing methods such as logistic regression model, artificial neural network model and decision tree model in terms of prediction accuracy. The model is also shown to have an advantage of providing several strategic information of practical use for consumer choice.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2008.10a
• /
• pp.1-7
• /
• 2008

The purpose of this paper is to develop a new classification method for predicting consumer choice based on genetic algorithm, and to validate its prediction power over existing methods. To serve this purpose, we propose a hybrid model, and discuss its methodological characteristics in comparison with other existing classification methods. Also, to assess the prediction power of the model, we conduct a series of experiments employing survey data of consumer choices of MP3 players. The results show that the suggested model in this paper is statistically superior to the existing methods such as logistic regression model, artificial neural network model and decision tree model in terms of prediction accuracy. The model is also shown to have an advantage of providing several strategic information of practical use for consumer choice.

• The Journal of Asian Finance, Economics and Business
• /
• v.7 no.2
• /
• pp.221-228
• /
• 2020

This study suggests an alternative to the conventional collaborative filtering method for predicting consumer choice, using case-based reasoning. The algorithm of case-based reasoning determines the similarity between the alternative sets that each subject chooses. Case-based reasoning uses the inverse of the normalized Euclidian distance as a similarity measurement. This normalized distance is calculated by the ratio of difference between each attribute level relative to the maximum range between the lowest and highest level. The alternative case-based reasoning based on similarity predicts a target subject's choice by applying the utility values of the subjects most similar to the target subject to calculate the utility of the profiles that the target subject chooses. This approach assumes that subjects who deliberate in a similar alternative set may have similar preferences for each attribute level in decision making. The result shows the similarity between comparable alternatives the consumers consider buying is a significant factor to predict the consumer choice. Also the interaction effect has a positive influence on the predictive accuracy. This implies the consumers who looked into the same alternatives can probably pick up the same product at the end. The suggested alternative requires fewer predictors than conjoint analysis for predicting customer choices.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.22 no.6
• /
• pp.157-167
• /
• 2023

The use of big data for transportation often involves using data that includes personal information, such as the driver's driving routes and coordinates. This study explores the creation of a route choice prediction model using a large dataset from mobile navigation apps using federated learning. This privacy-focused method used distributed computing and individual device usage. This study established preprocessing and analysis methods for driver data that can be used in route choice modeling and compared the performance and characteristics of widely used learning methods with federated learning methods. The performance of the model through federated learning did not show significantly superior results compared to previous models, but there was no substantial difference in the prediction accuracy. In conclusion, federated learning-based prediction models can be utilized appropriately in areas sensitive to privacy without requiring relatively high predictive accuracy, such as a driver's preferred route choice.

• Journal of the Korean Home Economics Association
• /
• v.46 no.5
• /
• pp.63-71
• /
• 2008

The purposes of this study were (1) to analyze the absolute and relative importance of tourists' choice attributes of restaurants and (2) to compare tourists' choice patterns with the prediction of restaurant operators in the tourism provinces of Gangwon-do, Empirical data for this study were collected from the 77 tourists and 66 restaurants operators. The attributes and attribute levels for the hypothetical profiles were decided from a focus group interview and 15 profiles were selected from fractional factorial designs. The SPSS/WIN 12,0 conjoint procedure was used to calculate the utility scores and simulate the profiles, According to the analysis on the relative importance of tourists' choice attributes of restaurants, food taste was the most important attribute(36.9%), followed by facility cleanliness(28.5%), dishes cleanliness(24.5%), price(19.3%) and service(18.3%). The tourists' ratings of the importance of the individual attributes differed from the ranking of the relative importance of the same attributes as derived from the conjoint analysis. The operators rated dishes cleanliness(27.6%) as also important, followed by food taste(27.7%), in choosing a restaurant Tourists' preference and operators' prediction of hypothetical profiles showed significant difference in L(p < .05), O(p < .01), M(p < .05), and H(p < .01) restaurants. Operators who want to differentiate themselves from their competitors should make decisions based on an increased understanding of tourists' restaurant choice decision process and measure the latent needs of tourists in the tourism provinces.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.24 no.4
• /
• pp.96-111
• /
• 1997

This study was carried out to identify users' choice behavior of theme parks. overland. Lotte World, Seoul Land, Dreamland and Children's Grand Park were selected as study areas. Both multinomial logic model(MNL), nested logic model(NMNL) and joint logit model wet$.$e test using a choice-based sample collected on study areas. Hausman-McFadden test showed that the MNL is not appropriate because the IIA assumption is violated. To avoid the problematic IIA assumption, the NMNL was tested. It splits similar alternatives into groups and nests separate decisions into hierarchical order to avoid the IIA assumption. Cluster analysis and discriminant analysis were conducted to find applicable nest structures. The inclusive value coefficient was 0.7788. It meant that sufficient condition of this model is met and users' choice behavior can be better understood by NMNL than MNL. The $\rho$2 value and accuracy of prediction of this model were 0.402 and 46.33% , respectively. Several comments were suggested to make the NMNL to be more reliable for future research on users' choice behavior of theme park.

• Communications for Statistical Applications and Methods
• /
• v.17 no.6
• /
• pp.811-827
• /
• 2010

This paper deals with a density estimation method in binary choice models that can be regarded as a statistical inverse problem. We use an orthogonal basis to estimate density function and consider the choice of an appropriate truncation parameter to reflect the model complexity and the prediction accuracy. We propose a data-dependent rule to choose the truncation parameter in the context of binary choice models. A numerical simulation is provided to illustrate the performance of the proposed method.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.28 no.2
• /
• pp.18-27
• /
• 2000

The importance of constraints has been one of major issues in recreation for prediction of choice behavior; however, traditional conjoint choice model did not consider the effects of these variables or fail to integrate them into choice model adequately. The purposes of this research are (a) to estimate the effects of constraints in theme park choice behavior by the constraints-induced conjoint choice model, and (b) to test additional explanatory power of the additional constraints in this suggested model against the more parsimonious traditional model. A leading polling agency was employed to select respondents. Both alternative generating and choice set generating fractional factorial design were conducted to meet the necessary and sufficient conditions for calibration of the constraints-induced conjoint choice model. Th alternative-specific model was calibrated. The log-likelihood ratio test revealed that suggested model was accepted in the favor of the traditional model, and the goodness-of-fit($\rho$$^2$) of suggested and traditional model was 0.48427 and 0.47950, respectively. There was no difference between traditional and suggested model in estimates of attribute levels of car and shuttle bus because alternatives were created to estimate the effects of constraints independently from mode related variables. Most parameters values of constraints had the expected sign and magnitude: the results reflected the characteristics of the theme parks, such as abundance of natural attractions and poor accessibility in Everland, location of major fun rides indoor in Lotte World, city park like characteristics of Dream Land, and traffic jams in Seoul. Instead of the multinomial logit model, the nested logit model is recommended for future researches because this model more reasonably reflects the real decision-making process in park choice. Development of new methodology too integrate this hierarchical decision-making into choice model is anticipated.

• Journal of the Korean Regional Science Association
• /
• v.11 no.1
• /
• pp.31-60
• /
• 1995

The distribution of economic activity over a mutually exclusive and exhaustive categorical industry-region matrix is modeled as a composition of two random components: the probability-like share distribution of jobs and the dynamic evolution of absolute aggregates. The former describes the individual activity location choice by comparing the predicted profitability of the current industry-region pair against that of all other alternatives based on the available information on industry-specific, region specific, or activity specific attributes. The latter describes the time evolution of macro-level aggregates using a dynamic reduced from model. With the seperation of micro choice behavior and macro dynamic aggregate constraint, the usual independence and identicality assumptions become consistent with the activity share distribution, hence multi-regional industrial migration can be represented by a set of probability evolution equations in a conservative Markovian from. We call this a Micro-Macro Composition Approach since the product of the aggregate prediction and the predicted activity share distribution gives the predicted activity distribution gives the predicted activity distribution which explicitly considers the underlying individual choice behavior. The model can be applied to interesting practical problems such as the plant location choice of multinational enterprise, the government industrial ploicy to attract international firms, and the optimal tax-transfer mix to influence activity location choice. We consider the latter as an example.