Optimization of rule-based systems in mHealth applications

Abstract mHealth applications are becoming more and more advanced, exhibiting capabilities to deliver innovative health services for improving the individual's comfort, enhancing the quality of life, promoting wellness and healthy lifestyle, or improving the adherence to therapies of remotely monitored patients. One of the most relevant components of such applications is represented by rule-based systems able both to reproduce deductive reasoning mechanisms and to explain how their outcomes have been achieved. Unfortunately, the efficiency of rule-based systems, especially on resource-limited mobile devices, rapidly decreases depending on the amount of data satisfying their rules as well as on the size and complexity of the whole rule base. Starting from these considerations, this paper proposes an optimization approach aimed at revising the structure of ontologies and rules built on the top of them that are contained into a rule-based system, with the goal of reducing the cost of evaluation for all its rules, by operating directly at the knowledge level. A general cost model is also presented to estimate the impact of research and identification of available rule instances to execute. Such a model is used to assess impacts and benefits due to the application of the proposed approach to a case study pertaining an mHealth app devised to evaluate eating habits of users in order to take under control their lifestyles and, thus, preserve their wellness. Finally, this theoretical evaluation is also transposed in a practical scenario, where the rule-based system embedded in the considered mHealth app is evaluated on a real smartphone, in terms of memory usage and overall response time. Moreover, a further study has been arranged in order to evaluate the impact of different rule conditions on the cost of evaluation of a knowledge base, and the eventual benefits drawn by their optimization. All the evaluation results show that the proposed approach offers an innovative and efficient solution to drastically reduce the cost of the evaluation of rule instances to execute and, thus, to build mHealth apps able to meet both real-time performance and computation intensive demands.

[1]  J. Carroll,et al.  Jena: implementing the semantic web recommendations , 2004, WWW Alt. '04.

[2]  Emily Gerth-Guyette,et al.  Impact of mHealth Chronic Disease Management on Treatment Adherence and Patient Outcomes: A Systematic Review , 2015, Journal of medical Internet research.

[3]  Daniel P. Miranker,et al.  Index support for rule activation , 1993, SIGMOD '93.

[4]  Charles L. Forgy,et al.  Rete: A Fast Algorithm for the Many Patterns/Many Objects Match Problem , 1982, Artif. Intell..

[5]  Toru Ishida,et al.  Optimizing Rules in Production System Programs , 1988, AAAI.

[6]  Donna J. Slovensky,et al.  mHealth: Transforming Healthcare , 2014 .

[7]  Murat Osman Ünalir,et al.  Reordering Query and Rule Patterns for Query Answering in a Rete-Based Inference Engine , 2005, WISE Workshops.

[8]  Giuseppe De Pietro,et al.  Design and validation of a light-weight reasoning system to support remote health monitoring applications , 2015, Eng. Appl. Artif. Intell..

[9]  Charles Lanny Forgy,et al.  On the efficient implementation of production systems. , 1979 .

[10]  Genichi Taguchi,et al.  Introduction to quality engineering.... , 2014 .

[11]  Shahriar Akter,et al.  Modelling the impact of mHealth service quality on satisfaction, continuance and quality of life , 2013 .

[12]  Daniel P. Miranker,et al.  On the Performance of Lazy Matching in Production Systems , 1990, AAAI.

[13]  Madhan Shridhar Phadke,et al.  Quality Engineering Using Robust Design , 1989 .

[14]  Nabil Arman.  Improving Rule Base Quality to Enhance Production Systems Performance , 2013 .

[15]  Wael Mustafa.  Optimization of Production Systems Using Genetic Algorithms , 2003, Int. J. Comput. Intell. Appl..

[16]  Enrico Motta,et al.  How Much Semantic Data on Small Devices? , 2010, EKAW.

[17]  Daniel P. Miranker.  TREAT: A new and efficient match algorithm for AI production systems , 1988 .

[18]  Peter Van Weert,et al.  Efficient Lazy Evaluation of Rule-Based Programs , 2010, IEEE Transactions on Knowledge and Data Engineering.

[19]  Emily Knight,et al.  Health Promotion Through Primary Care: Enhancing Self-Management With Activity Prescription and mHealth , 2014, The Physician and sportsmedicine.

[20]  Murat Osman Ünalir,et al.  Optimizing a Rete-based Inference Engine using a Hybrid Heuristic and Pyramid based Indexes on Ontological Data , 2007, J. Comput..