A rule based expert system to advise on air-filtering plants for indoor spaces in UAE

Doshi, G.J., Stevens, J.D. and Khan, S.T. (2019) A rule based expert system to advise on air-filtering plants for indoor spaces in UAE. In: 5th International Conference on System Control and Communications Proceedings, 21 - 23 December 2019, Wuhan, China

Abstract

The purpose of this study was the development of an expert system that is able to recommend types of plants for the removal of toxic substances in a given artificial ecosystem. The domain was restricted to office environments within the United Arab Emirates (UAE).

A literature review revealed a significant gap in research related to systems that help select plants based on a given environment. This eventually led to the creation of the Plant Recommender Expert System (PRES).

PRES utilizes a set of inputs (for example expected humidity, average temperature, light quality, etc.). The system will then recommend the type(s) of plants that should be purchased to suit the given environment conditions.

Horticultural experts usually give recommendations for these types of problem domains but such an expert may not always be available, and so the concept here is to encode the expertise knowledge in an intelligent system to ensure uninterrupted availability of the expert knowledge.

The system was evaluated using several case studies with known outcomes. The PRES suggested plants for environments that were described via a set of inputs in these trials.

The initial phase, being a prototype of PRES was created with the aim of helping non-expert users leverage the natural air-detoxification properties of plants.

While in its current configuration the system is not capable of learning, it can at a later stage be integrated with other AI methods such as the use of decision trees to create further rules or data driven approaches such as the use of Neural Networks for the classification of plants relevant to a given domain. However, the latter approach will first necessitate the creation of a relevant data set.

Item Type:	Conference Paper
Conference Website:	http://www.icscc.org/
URI:	http://researchrepository.murdoch.edu.au/id/eprint/54963

Item Control Page

Downloads