Towards an improved internet of things sensors data quality for a smart aquaponics system yield prediction.

Autor:	Eneh AH; Department of Computer Science, University of Nigeria, Nigeria., Udanor CN; Department of Computer Science, University of Nigeria, Nigeria., Ossai NI; Department of Zoology & Environmental Biology, University of Nigeria, Nigeria., Aneke SO; Department of Computer Science, University of Nigeria, Nigeria., Ugwoke PO; Digital Bridge Institute, Nigeria Communications Commission, Abuja, Nigeria., Obayi AA; Department of Computer Science, University of Nigeria, Nigeria., Ugwuishiwu CH; Department of Computer Science, University of Nigeria, Nigeria., Okereke GE; Department of Computer Science, University of Nigeria, Nigeria.
Jazyk:	angličtina
Zdroj:	MethodsX [MethodsX] 2023 Oct 11; Vol. 11, pp. 102436. Date of Electronic Publication: 2023 Oct 11 (Print Publication: 2023).
DOI:	10.1016/j.mex.2023.102436
Abstrakt:	The mobile aquaponics system is a sustainable integrated aquaculture-crop production system in which wastewater from fish ponds are utilized in crop production, filtered, and returned for aquaculture uses. This process ensures the optimization of water and nutrients as well as the simultaneous production of fish and crops in portable homestead models. The Lack of datasets and documentations on monitoring growth parameters in Sub-Saharan Africa hamper the effective management and prediction of yields. Water quality impacts the fish growth rate, feed consumption, and general well-being irrespective of the system. This research presents an improvement on the IoT water quality sensor system earlier developed in a previous study in carried out in conjunction with two local catfish farmers. The improved system produced datasets that when trained using several machine learning algorithms achieved a test RMSE score of 0.6140 against 1.0128 from the old system for fish length prediction using Decision Tree Regressor. Further testing with the XGBoost Regressor achieved a test RMSE score of 7.0192 for fish weight prediction from the initial IoT dataset and 0.7793 from the improved IoT dataset. Both systems achieved a prediction accuracy of 99%. These evaluations clearly show that the improved system outperformed the initial one.•The discovery and use of improved IoT pond water quality sensors.•Development of machine learning models to evaluate the methods.•Testing of the datasets from the two methods using the machine learning models. Competing Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. (© 2023 The Authors. Published by Elsevier B.V.)
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu