ارتباط شبکه های بیزی و مدل سازی مسیر PLS برای تجزیه و تحلیل علت و معلولی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|28999||2010||6 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Expert Systems with Applications, Volume 37, Issue 1, January 2010, Pages 134–139
Causal knowledge based on causal analysis can advance the quality of decision-making and thereby facilitate a process of transforming strategic objectives into effective actions. Several creditable studies have emphasized the usefulness of causal analysis techniques. Partial least squares (PLS) path modeling is one of several popular causal analysis techniques. However, one difficulty often faced when we commence research is that the causal direction is unknown due to the lack of background knowledge. To solve this difficulty, this paper proposes a method that links the Bayesian network and PLS path modeling for causal analysis. An empirical study is presented to illustrate the application of the proposed method. Based on the findings of this study, conclusions and implications for management are discussed.
In recent years, knowledge management and related strategy concepts are often promoted as important components of organizations’ survival strategies (Martensson, 2000). Knowledge management is regarded as a key source of sustainable competitive advantage (Holsapple and Singh, 2001 and Liao, 2003), and is also seen as playing a fundamental role in the process of transforming individual knowledge into organizational knowledge (Liebowitz, 2001). Knowledge is generally considered as intangible, and is difficult to measure, but it is generally accepted that it sometimes increases through use (Wiig, Hoog, & Spek, 1997). More importantly, for the purpose of advancing the quality of decision-making and thereby facilitating a process of transforming strategic objectives into effective actions, causal knowledge based on causal analysis is needed (Lin and Wu, 2008, Nadkarni and Shenoy, 2004 and Tan and Platts, 2003). Several creditable studies have placed emphasis on the usefulness of causal analysis techniques. Tan and Platts (2003) conduct an appraisal of causal analysis techniques ranging from fishbone (Ishikawa) diagrams, Why/Why diagrams, influence diagrams, mind maps, to cognitive maps. In an especially relevant study, Lin and Wu (2008) implement a fuzzy DEMATEL method to produce a causal diagram. In addition, Bayesian networks and PLS path modeling are popular causal analysis techniques. Through these causal analysis techniques, causal maps can be created. Causal maps represent the causal knowledge of subjects in a specific domain, and they have been applied widely in the areas of policy analysis and management sciences to demonstrate the relationships between relevant factors, knowledge, and conditions (Nadkarni & Shenoy, 2004). Among causal analysis techniques, the PLS path modeling is particularly famous for its successful applications in customer satisfaction analysis; both the American Customer Satisfaction Index (ACSI) and the European Customer Satisfaction Index (ECSI) were constructed using PLS path modeling. However, many researchers and experts have experienced a certain amount of difficulty regarding how to establish the causal directions between constructs, due to lack of background knowledge or previous theoretical support. To deal with this difficulty, this paper proposes using the Bayesian network prior to implementing PLS path modeling for causal analysis. The Bayesian network is a causal map, a kind of graphical representation of an expert’s knowledge, based on probability theory (Nadkarni & Shenoy, 2004). The Bayesian network enjoys the advantage that it needs no rigid statistical assumptions: it graphically displays as a directed acyclic graph (DAG), and represents a set of conditional independence constraints among a given number of variables and their related conditional probability distributions (Lauria & Duchessi, 2007). Because of the special nature and merit of the Bayesian network, the DAG can serve as a guide to help us decide the causal directions between constructs when using PLS path modeling. This paper, therefore, suggests a method that links the Bayesian network and PLS path modeling for causal analysis. An empirical study is presented to illustrate the application of the proposed method. The remainder of this paper is organized as follows. In Section 2, the proposed method is presented. In Section 3, the Bayesian network and PLS path modeling are discussed. In Section 4, an empirical study is presented by way of illustration. Finally, based on the findings of this research, conclusions and implications for management are presented.
نتیجه گیری انگلیسی
Profoundly reasoned decision-making relies greatly on the ability of decision-makers to model complex cause–effect relationships, and thus to undertake constructive actions based on dependable causal knowledge. Causal knowledge based on causal analysis can improve the quality of decision-making and facilitate the transformation and conversion of strategic objectives into effective actions. Several studies have suggested various causal analysis techniques to achieve causal knowledge. Among causal analysis techniques, PLS path modeling is one of the most powerful and popular causal analysis techniques. However, few studies have concerned themselves with the issue of theory dependency in implementing PLS path modeling. When using PLS path modeling in a situation lacking the support of an underlying previous theory, serious problems may arise. This is because the task of deciding the causal direction between constructs tends to become so complicated so we are compelled to guess at enormous numbers of possibilities for causal directions. In an effort to solve this difficulty, this paper proposes an effective method of pursuing a causal analysis process, i.e. the author suggests implementing the Bayesian network prior to conducting PLS path modeling. Additionally, an empirical study is presented to demonstrate the successful application of the proposed method. The proposed method embodies a successful combination of the Bayesian network with the TAN search algorithm and PLS path modeling for causal analysis. Although this methodology is successful in that the task of deciding the causal directions between attributes becomes easy through use of the proposed method, this study has some limitations. For instance, although the proposed method can save us from countless instances of trial and error in deciding causal directions, it is not necessarily the best solution when using the TAN search algorithm to settle on the causal directions. This is a significant issue calling for further future research.