دانلود رایگان مقاله انگلیسی توسعه سیستم خبره پوششی: تعاریف، مدل ها و چالش هایی برای آینده - امرالد 2017

Structured abstract

Purpose A Wearable Expert System (WES) is an expert system designed and implemented to obtain input from and give outputs to wearable devices. Among its distinguishing features are the direct cooperation between domain experts and users, and the interaction with a knowledge maintenance system devoted to dynamically update the knowledge base taking care of the evolving scenario. Design/Methodology approach The WES development method is based on the KAFKA framework. KAFKA employs multiple Knowledge Artifacts, each devoted to the acquisition and management of a specific kind of knowledge. The KAFKA framework is introduced from both the conceptual and computational points of view. An example is given which demonstrates the interaction, within this framework, of Taxonomies, Bayesian Networks and Rule Based Systems. An experimental assessment of the framework usability is also given. Findings The most interesting characteristic of WESs is their capability to evolve over time, due both to the measurement of new values for input variables and to the detection of new input events, that can be used to modify, extend and maintain knowledge-bases and to represent domains characterized by variability over time. Originality/value Wearable Expert System is a new and challenging concept, dealing with the possibility for a user to develop his/her own decision support systems and update them according to new events when they arise from the environment. The system fully supports domain experts and users with no particular skills in knowledge engineering methodologies, to create, maintain and exploit their expert systems, everywhere and when necessary.

7 Conclusion and Future Works

This paper has discussed the design of Wearable Expert Systems. The proposed design methodology is KAFKA, a framework based on the Knowledge Artifact conceptual model, which is general enough to be adopted in different contexts and programming paradigms.

From the conceptual point of view, KAFKA aims at making the development of knowledge-based systems (in particular, rule-based systems) quicker and simpler through the reduction of knowledge engineer responsibilities. In this way, the knowledge engineering process is focused on the kinds of knowledge involved in the decision making activity rather than on how to model it, representing a radical change of perspective if compared with classical approaches like CommonKADS and MIKE. In this sense, KAFKA philosophy is closer to methodologies like MOKA (Stokes et al., 2001) and KNOMAD (Curran et al., 2010), proposed in the knowledge-based engineering field as product-oriented (Verhagen et al., 2012) rather than process-oriented tools for supporting users in the configuration of objects.

The WESs which can be developed employing this methodology bear some distinguishing features with respect to traditional expert systems. The WES knowledge base may change dynamically, following the long-term evolution of the monitored system and of its surrounding environment. Moreover, the presence of a centralized knowledge maintenance system, in principle common to a large number of WES instances, permits to exploit the massive amount of information coming from this large set of wearable devices: for instance, in the example of section 5, the Queue status may be monitored reliably by collecting data from all the application users in the city, tens or maybe hundreds of devices.