Asian J. of Management Science and Applications (7 papers in press)
An Analytical Model of Website Relationships Based on Browsing History Embedding Considerations of Page Transitions
by Taiju Hosaka, Haruka Yamashita, Masayuki Goto
Abstract: In recent years, obtaining a large amount of information and receiving
various services through websites has become possible. Consequently, web
browsing activities are increasing exponentially, and numerous companies are
conducting their businesses online. In this scenario, usersweb browsing behavior
containing their detailed activities is one of the most important topics in web
marketing analysis. Several studies have been conducted with a focus on analyzing
the web browsing history of users. Due to the assumption that a users browsing
history reflects his/her preferences, understanding these preferences is crucial
for optimizing the marketing measures. Distributed representation models can
effectively represent the characteristics of websites based on the structure of the
browsing history data. These methods can flexibly represent the relationships
between the objects by embedding objects with strong relationships into an
embedding space as similar representations. However, a large amount of browsing
behavior without a clear browsing purpose may sometimes be present in the
history data. Therefore, learning representations that reflect the overall browsing
behavior is undesirable from the perspective of robustness. In this paper, we
propose the sparse skip gram model derived from the sparse continuous bagof-
words (CBoW) model using the regularized online learning approach for
analyzingword meanings, based on which the web browsing behaviors of the users
can be analyzed on an embedding space. In addition,weapply our method to actual
browsing history data and discuss the findings acquired from the analytical results.
We show that our proposed model represents the characteristics of websites with
subspaces in the embedding space.
Keywords: Distributed representation; Word2Vec; Sparse regularization; Follow-the-Regularized-Leader; Data Mining; Business analytics; Web marketing; Semantic analysis; Interest detection; Browsing history data.
A Study on Recommender System Considering Diversity of Items Based on LDA
by Zhiying Zhang, Taiju Hosaka, Haruka Yamashita, Masayuki Goto
Abstract: With the rapid development of information technology, a recommender system making use of users behavior data, such as browsing history or ratings for items, is now one of the important tools for searching contents or products. Recently, it has been shown that diversifying the recommendation lists in recommender systems could satisfy users potential needs. In a previous research, the diversity of recommender system can be raised by the topic diversification method based on Latent Dirichlet Allocation (LDA); however, since the items belonging to the same topic are not diversified, the recommended items in the list shown to a user tend to be similar. Therefore, this research proposes a method for a recommendation system that diversifies items in each topic based on topic information obtained by LDA. Experimental results with MovieLens datasets demonstrate that our approach keeps accuracy of the recommendation and realizes more diversified recommendation.
Keywords: Recommender System; LDA; Latent Dirichlet Allocation; Topic Model; Machine Learning; Diversity.
Multi-period Stochastic Lateral Transshipment Problem for Rental products
by Koji Aragane, Tomoki Fukuba, Takayuki Shiina
Abstract: In this study, we employed a multi-period stochastic programming
model to examine the lateral transshipment problem of rental products and verified
the effectiveness of this model under several assumptions about demand trends.
The design and operation of multi-period supply chain networks is a large-scale
optimization problem under uncertainty. Lateral transshipment between bases is
effective in improving service levels and in reducing inventory. In this study, we
assume that the quantity of demand and length of rentals follow a probability
distribution (since it deals with rental products). We construct a scenario tree to
consider the uncertainty of rental products in multi-periods using the K-means
method. Considering demand trends, we construct three types of scenario trees
that represent flat, growth, and decline trends. We demonstrate the effectiveness
of the model for each trend through numerical experimentation.
Keywords: Stochastic programming model; Lateral transshipment; rentalrnproduct; Supply chain; K-means method; Scenario tree; Multi-periods.
Sourcing decision with capacity reservation under supply disruption risk
by Kotomichi Matsuno, Jiahua Weng, Xianghe Shao
Abstract: Over the last decade, more than 50% of organizations in the world have annually experienced a supply disruption such as geopolitical instability and transportation failure. Therefore, it has become crucial for manufacturers to comprehensively evaluate sourcing decisions not only on procurement costs, quality, and delivery but also on disruption risk. Although risk-hedging on supply disruption could be realized with evaluations on suppliers, there are no reduction effects on immense opportunity losses by manufacturers due to supply disruption. Owing to disasters such as Hurricane Katrina 2005, Great East Japan Earthquake 2011, Thailand Flood 2011, and Kumamoto Earthquake 2016, countless manufacturers have suffered from heavy losses in component procurement. In this study, as an opportunity loss mitigation countermeasure, a sourcing decision with capacity reservation (CR) is proposed for manufacturers. The effectiveness of the CR model is discussed by comparing it with the supplier selection and order allocation model without CR by disruption probability and supply demand relationship.
Keywords: Disruption Risk; Sourcing Decision; Capacity Reservation; Procurement; Supply Chain Contract.
A New Approach on the Lowest Cost Problem in Data Envelopment Analysis
by Xu Wang, Kuan Lu, Takashi Hasuike
Abstract: This paper aims at solving the lowest cost problem in Data Envelopment Analysis(DEA), which is to provide an efficient target for an inefficient decision making unit(DMU) with the lowest adjustment costs. For this purpose, a new approach based on the least distance DEA model is proposed. Here, the marginal costs of adjusting the inputs and outputs are assumed to be known and symmetrical. For the practical merit, different with the existing studies, our approach is able to increase inputs and decrease outputs. Numerical experiments are conducted to compare the performance of the proposed approach with previous existing studies. The results show that the proposed approach can always provide an efficient target with no higher total adjustment costs than the costs of targets provided by previous approaches. Thus, the proposed approach is more practical and useful.
Keywords: data envelopment analysis; lowest cost problem; least distance model; symmetrical marginal costs.
Building a Performance Indicator to Investigate the Robustness of Water Supply Utilities in Japan
by Yuji Kawase, Yusuke Maeno, Kumudu Piyasena, Tatsuo Oyama
Abstract: In this study, we build a work performance indicator (WPI) for quantitatively measuring the performances of water supply utilities in Japan. The WPI is defined from three different perspectives: management, facility/equipment, and operations. It can be used to measure the robustness of a water supply network system. Based on actual data expressing the work performances of Japanese water supply utilities from 1980 to 2018, we illustrate the numerical results of the WPIs, so as to determine their regional characteristics. We apply statistical data analysis techniques (such as a cluster analysis and principal component analysis) to calculate the WPIs, and investigate the historical and regional characteristics and trends of water supply utilities in Japan during the investigated period. These approaches can be applied to design and improve natural disaster mitigation policies, such as those focusing on earthquakes in Japan.
Keywords: water supply utility; work performance indicator; robustness; water supply network; statistical data analysis; cluster analysis; principal component analysis.
Organisational Culture and Shared Leadership in Chinese Enterprises
by Cong Xu
Abstract: The current study investigates the relationship between performance-based organisational culture (i.e., innovation-oriented culture, outcome-oriented culture, and detail-oriented culture) and three dimensions of shared leadership (i.e., collective-achievement leadership, mutual-engagement leadership, and cohesive-support leadership). Multiple regression results based on data obtained from 12 Chinese firms demonstrated that innovation-oriented culture was significantly related to collective-achievement leadership and cohesive-support leadership; outcome-oriented culture was significantly related to mutual-engagement leadership and cohesive-support leadership; and attention to detail was significantly related to mutual-engagement leadership. It was also hypothesised that there would be moderating effects of firm ownership and location in Chinese markets. However, the moderating effects were not as strong as it was expected. The results provide novel insights into theories and practise of shared leadership and organisational culture in China.
Keywords: shared leadership; organisational culture; Chinese markets; firm ownership; firm location