Ordinal logistic models are widely used in various fields of research. Among ordinal logistic models, this paper adopts continuation ratio logistic models. Since the models are too flexible, the estimates of parameters in the models tend to be unstable. For the reason, it is often assumed that the values of coefficient parameters that same variables have are equal. The assumption is called “equal slope assumption”. Yasukawa & Tsubaki (2003) used the information criterion AIC for determining which the model satisfies equal slope assumption or not. However, as the number of covariates increases, the number of models to be selected increases exponentially. Therefore, model selection procedures with information criteria are realistically impossible.
 To overcome the problem, we propose a sparse ordinal logistic model (SOL), which is constructed by combining the fused lasso (Tibshirani et al., 2005) with continuation ratio logistic models. Monte Carlo simulations and a real data analysis are performed to investigate the efficiency of SOL.

View full abstract

 Falling is a common health problem for the elderly. “Bedridden elderly persons” is a medical and social problem, and it is mainly caused by falling. Conventional fall detection systems have various limitations such as privacy. In this study, we investigate the possibility of analyzing fall detection system based on the microwave Doppler radar. We developed a physical model to represent the Doppler signatures of falling and applied it to the fall detection system. To demonstrate the effectiveness of the proposed method, we conducted experiments to assess the accuracy of the fall detection. Results indicate that high accuracy was achieved when the method was implemented for fall detection system.

View full abstract

 In this paper, we suggest some strategies for reducing waiting time in order to improve the quality of services provided by a ward office. Actual data of a reception guide system provided by Itabashi City in Tokyo were statistically analyzed, from which two strategies were proposed. Based on this data, we constructed simulation models of the two strategies. From the results of computer simulations of customer behavior, an increase in the workforce for the counter with the largest number of customers is more effective than rearranging the counters in reducing waiting time.

View full abstract

 In this study, we propose a method of reducing overall waiting time for queueing system in a public authority. The system has a common queue with two different services. The proposed method is based on the priority processing of the jobs with the shortest processing times regardless of order of arrival and reducing the overall waiting time. We assume that the arrivals to the queue follow a Poisson distribution and that the service time for the two services is exponentially distributed, but each service has a different mean value. To demonstrate the effectiveness of the proposed method, we carried out a simulation that assessed the performance of the queueing control. Simulation results indicate that high performance was achieved when the method was implemented for queueing system.

View full abstract

 In this study, we apply cascading failures to merchandise networks and propose a model that expresses the transition of consumers' want due to sellout. Its efficacy is verified using the POS data from supermarkets. Although conventional models consider only the material flow in a network, this model enables us to consider the flow of latent traits in a network. In detail, we first reconstruct the network in which a sellout product disappears, and then describe consumers' want as betweenness centrality. Secondly, we focus on products influenced by the sellout and categorize them based on how they are related to the sellout. Finally, in order to varidate the model, we also examine whether the products influenced by the sellout will actually be sold out.

View full abstract

 In sports like soccer, with lots of player and ball movement in the game, spectators usually only roughly brasp the flow of the game as a whole, and focus specifically on the characteristics of the team.
During a game, we may basically understand team ball possession by physically watching; however, we can understand it logically by analyzing play data during the game. Constructing an application for such a purpose, easier visualization of the overall game flow, by calculating possession rates from soccer play data, would be extremely helpful; and RStudio would be a good tool. Thus, we made a visualization application that can easily measure many games by using manipulate function and Shiny of RStudio.

View full abstract