Duk Bin Jun

A Choice-Based Diffusion Model for Multiple Generations of Products

Abstract We incorporate diffusion effects and choice effects in an integrated model to capture simultaneously the diffusion and substitution processes for each successive generation of a durable technology. The choice literature generally ignores demand dynamics and previous multigeneration diffusion models rarely include control variables. The proposed model is a combination of the two approaches. The basic premise of the proposed model states that the replacement of an older product by a newer one is based on the choice behavior of consumers, where consumers choose a product to maximize their utility. Then we can derive the implied relationships among choice probabilities, diffusion processes, and marketing mix variables. To verify the proposed model, we also analyze the IBM mainframe market and worldwide DRAM (dynamic random access memory) market.

Forecasting telecommunication service subscribers in substitutive and competitive environments

Abstract The telecommunications market is expanding rapidly and becoming more substitutive and competitive. In this environment, demand forecasting is very difficult, yet important for both practitioners and researchers. In this paper, we adopt the modeling approach proposed by Jun and Park [Technological Forecasting and Social Change 61 (1999)]. The basic premise is that demand patterns result from choice behavior, where customers choose a product to maximize their utility. We apply a choice-based substitutive diffusion model to the Korean mobile telecommunication service market where digital service has completely replaced analog service. A choice-based competitive diffusion model is also formulated and applied to the case where two digital services compete. In comparison with Bass-type models, these two models provide superior fitting and forecasting performance. Finally, we suggest a new choice-based diffusion model to describe an environment in which substitution and competition occur simultaneously and show the application results. The choice-based model is useful in that it enables the description of such complicated environments and provides the flexibility to include marketing mix variables such as price and advertising in the regression analysis.

Forecasting demand for low earth orbit mobile satellite service in Korea

Forecasting a new service diffusion process is critical in designing marketing strategies and analyzing the costs and benefits for service providers. It is very difficult, however, in cases that data are not available. We suggest the combination of analogy and survey to forecast the demand for Low Earth Orbit (LEO) mobile satellite service in Korea. First, we analyze the diffusion of existing mobile phone service, which is similar to LEO service. The diffusion parameters for mobile phone service are then used in a model for LEO service. A survey was made on two hundred fifty‐five subscribers of existing mobile phone service in Korea. We estimate the potential market size of LEO service by applying the logit model to the survey data. Then, we forecast the annual demand for LEO service in Korea from 1998 to 2005. We also derive the price elasticity of market potential of LEO service.

State space trend‐cycle decomposition of the ARIMA(1,1,1) process

It has been common practice to decompose an integrated time series into a random walk trend and a stationary cycle using the state space model. Application of state space trend-cycle decomposition, however, often results in a misleading interpretation of the model, especially when the observability of the state space model and the redundant relationships among the model parameters are not properly considered. In this study, it is shown that spurious trend-cycle decomposition, discussed by Nelson (1988), results from an unobservable state space model, and the usual assumption of independent noise processes in the model results in parameter redundancy. Equivalence relationships for the ARIMA(1,1,1) process and the state space model consisting of a random walk trend and an AR(1) cycle, where the noise processes of the trend and of the cycle are generally correlated, are also derived.

Growth-cycle decomposition diffusion model

The diffusion model has been widely used to explain the S-shaped cumulative growth of markets for retail service and consumer durable goods. In many situations, sales fluctuate because of both growth of innovation diffusion and transitory changes in an external factor, called cycle. The traditional diffusion model, however, cannot distinguish between the two. We develop the growth-cycle decomposition diffusion model to distinguish the growth from the cycle, where the cycle incorporates a few external variables determining the transitory sales environment. The proposed model is applied to estimating the diffusion process for the annual sales of room air conditioners in Korea.

Long Term Mean Reversion of Stock Prices Based on Fractional Integration

In this study we examine the long term behavior of stock returns. The analysis reveals that negative autocorrelations of the returns exist for a super-long horizon as long as 10 years. This pattern, however, contrasts to predictions of previous stock price models which include random walks. We suggest the introduction of a fractionally integrated process into a nonstationary component of stock prices, and demonstrate empirically the existence of the process in NYSE stock returns. The predicted values of autocorrelation from our stock price model confirm the super-long term behavior of the returns observed in regression, indicating that inefficiency in the stock market could remain for a long time.

A Bayesian DYMIMIC Model for Forecasting Movie Viewers

In this study, we explore the issue of how to enhance forecast of the box office sales, an all-time question for managers in the motion picture industry. The conceptual core of our approach is the expected sales. The expected sales of agents in the movie market (i.e. screen managers at supply side or potential moviegoers at demand side) play important role in predicting the actual sales. We pay attention to the uniqueness of the expected sales; it is latent and evolving over time. This leads to a quick sense that incorporating these components into the model is a natural choice and thus critical for proper forecast of the movie for future period. Based on this notion, we proposed a simple DYMIMIC model for forecasting the box office sales. The model based on a simplified intuitive story of movie consumption behaviors, spontaneous demand and socially driven demand, was calibrated and tested on the actual movie data in the United States and other countries. The model allows for evolution of the latent expectation in the simplest way, and it captures both the cross-sectional unobserved heterogeneity across countries and the effects of the sequential releases over countries. Compared to previous forecasting models, the suggested approach offers a simple yet informative platform of the model that one can add variables such as movie attributes and marketing activities.

Classification, relationships and forecasting models for telecommunication services

In forecasting future market size for telecommunication services, it is customary to analyze individually the diffusion processes of the services that make up the whole market and aggregate them. Recently, however, many existing telecommunication services have diversified and new services have arisen in an effort to satisfy customer needs. Thus, an aggregated forecast should consider the various relationships among telecommunication services, such as competitiveness and complementariness in view of the customers desires for telecommunication. In this paper, a framework for classifying telecommunication services is proposed and independent, competitive and complementary relationships are defined according to customer needs, customer premise equipment, cost and network. Forecasting models based on such relationships are applied to telecommunication services in Korea.

Copula-Based Simultaneous Approach to Multivariate Alternative Choice and Quantity Choice

This paper aims to examine correlations in shopping situations. First, there is a certain amount of correlation between alternative choices. Specifically, the alternatives from different categories but from a same brand might be purchased together. Second, alternative choice and quantity choice could be correlated each other. A consumer tends to purchase tooth paste with large amount, but hand cream with small amount. Third, quantity choices could be correlated each other. The purchased quantity of fabric softener should depend on the purchased quantity of laundry detergent. To explain these correlations, the model must deal with multivariate incidence and quantity outcomes. Therefore, we developed a new copula-based approach to simultaneously deal with them, so that it could directly control and capture the correlations. Also, we found that if the copula function is a multivariate-FGM copula, then the likelihood is closed form that is easy to estimate. We apply this model to IRI scanner panel data and estimate the model by using Bayesian method. In this data set, we could find strong dependencies between alternative choices, between alternative choice and quantity choice, and between quantity choices. In addition, more efficient promotion strategy of two products from a same brand but different categories is drawn from our model.

Forecasting a daily time series with varying seasonalities: an application to daily visitors to farmland in Korea

Abstract An accurate forecast of a daily time series often plays an important role in many managerial and industrial decisions related to production planning, scheduling and control. The fluctuation in daily time series are affected not only by the quarterly, monthly and weekly seasonalities, but also by both solar and lunar holidays in Asian culture. However, because the holidays make the seasonal factor irregular any single traditional seasonal model fails to describe the complicated relations among the various seasonalities and the changing solar and lunar holiday effects. In this study, we develop a daily index which incorporates into a single measure the effects of quarterly, monthly and weekly seasonalities and the effects of holidays. A time series model is also developed to forecast a daily time series by using the developed daily index with an application to the daily number of visitors to a large public amusement park in Korea. The results of the proposed models application are compared with those of the ARIMA model and the regression model.