Article Search
닫기

Original Article

Split Viewer

International Journal of Fuzzy Logic and Intelligent Systems 2023; 23(2): 214-228

Published online June 25, 2023

https://doi.org/10.5391/IJFIS.2023.23.2.214

© The Korean Institute of Intelligent Systems

Prediction of Soil Quality in Rwanda for Ideal Cultivation of Potato () Using Fuzzy Logic and Machine Learning

Christine Musanase1,2, Anthony Vodacek2,3 , Damien Hanyurwimfura1,2 , Alfred Uwitonze1,2 , Aloys Fashaho1,2 , and Adrien Turamyemyirijuru1,2

1African Center of Excellence in Internet of Things, University of Rwanda, Kigali, Rwanda
2Chester F. Carlson Center for Imaging Science, Rochester Institute of Technology, Rochester, NY, USA
3College of Agriculture, Animal Sciences and Veterinary Medicine, University of Rwanda, Musanze, Rwanda

Correspondence to :
Christine Musanase (musanasechristine@gmail.com)

Received: March 23, 2023; Revised: May 23, 2023; Accepted: June 15, 2023

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

The ability to estimate soil quality has great value for agriculture, especially for low-income regions with minimal agricultural and financial resources. This prediction provides users with information that is useful in determining whether the soil is suitable for a specific crop, such as potato (Solanum tuberosum). Farmers in Rwanda lack information on soil quality. There are not enough soil laboratories to perform the requisite measurements of NPK, pH, and organic carbon, nor are there enough experts to analyze the data and provide farmers with timely results. The prime objective of the proposed study is to develop a predictive framework that can estimate soil quality for the ideal cultivation of potato (Solanum tuberosum) considering a case study of Rwanda. In this study, bootstrapping is used to augment the small soil dataset, and fuzzy logic is used to label soil data into four classes of soil suitability, with verification of the labeling by soil experts. Several machine learning methods are then tested on the labeled data, resulting in the classification of suitability for the augmented dataset and an assessment of their performance as a way to support experts in predicting soil quality. All machine learning methods applied were viable, with the best performance achieved using an artificial neural network. The quantified outcome showed that the adoption of a neural-network-based scheme has an average accuracy of 32% in contrast to other learning schemes. However, 70%-80% accuracy was achieved upon the adoption of fuzzy logic.

Keywords: Soil quality, Fuzzy logic, Artificial intelligence, Rwanda, Machine learning, NPK, Predictive model

This research received no external funding.

The authors declare no conflicts of interest.

Christine Musanase is currently serving as an assistant lecturer in the Department of Information Systems, the School of ICT at the University of Rwanda, College of Science and Technology. She holds a bachelor’s degree in information technology from University of Rwanda and master’s degree in information systems from the University of Rwanda. She is a program leader of Information Systems. She is pursuing her Ph.D. in research in wireless intelligence sensor networks at the University of Rwanda through the African Center of Excellence in the Internet of Things (Rwanda). She is a member of RAWISE and OWSD National chapter. She is certified as an Oracle Database Developer & Programmer, Rapid Prototyping for Internet of Things, Python programming, Data Science, and R Programming for Data Analytics. Her research interests are informatics, wireless intelligence sensor networks, Internet of Things, artificial intelligence, data analytics, and machine learning. She is a member of three research grant projects titled (1) Tools for Evaluating African Lakes, (2) IoT and AI Applied Research Results Commercialization through the Incubation Hub, and (3) IoT Empowered Precision Agricultural Techniques for Improved Rice Production: An Automated Irrigation and Fertilization Application System for Small-scale Rice Producers in Rwanda. E-mail: musanasechristine@gmail.com

Anthony Vodacek is a full professor of Imaging Science at Rochester Institute of Technology (RIT). He received his B.S. (Chemistry) in 1981 from the University of Wisconsin-Madison, and his M.S. and Ph.D. (environmental engineering) in 1985 and 1990, respectively, from Cornell University. His areas of research lie broadly in multimodal remote sensing, with a focus on the coupling of imaging with modeling for monitoring human and natural terrestrial and aquatic systems. His expertise is in spectral phenomenology, image interpretation, machine learning, and dynamic data-driven application systems. He has recently applied these methods to projects addressing vehicle tracking, precision agriculture, and harmful algal blooms. His newest research areas involve remote sensing of the African Great Lakes and remote sensing of insects in the context of biodiversity assessment. He has worked in Rwanda for more than a decade on various teaching and research projects. Vodacek is on the Fulbright Specialist roster (2018–2023), is an Associate Editor for the Journal of Great Lakes Research, is a Senior Member of IEEE, supports the IEEE Geoscience and Remote Sensing Society global initiative as an ad hoc regional liaison to Sub-Saharan Africa, and is a Corresponding Fellow of the Pan-African Scientific Research Council. E-mail: axvpci@rit.edu

Damien Hanyurwimfura is an associate professor and the Acting Director of the African Center of Excellence in Internet of Things (ACEIoT), College of Science and Technology, University of Rwanda. He received his Bachelor of Engineering degree in computer engineering and information technology from the University of Rwanda (formerly KIST) in 2005. He obtained his Master of Engineering degree in computer science and technology and Ph.D. degree in computer science and technology from Hunan University, China, in 2010 and 2015, respectively. He has served as the Head of PhD studies and Research in the ACEIoT for four years. He has published and co-authored over 30 research papers in leading international journals and conferences. He participated in many AI workshops as a speaker. He has secured four research grants at the national and regional levels as a principal investigator or co-PI. His research interests include most aspects of data mining, machine learning, computer security, watermarking, Internet of Things, hate speech detection, and recommender systems. E-mail: hadamfr@gmail.com

Alfred Uwitonze is a senior lecturer & Dean of the School of Information and Communication Technology at University of Rwanda, College of Science and Technology. He received a Bachelor of Science degree in electronics and telecommunication engineering from the University of Rwanda (UR), College of Science and Technology, Rwanda, in 2005 and MSc degree in communication and information systems from Huazhong University of Science and Technology (HUST), China, in 2009. He completed his Ph.D. in information and communication engineering at the Huazhong University of Science and Technology (HUST), China, in 2017. His Ph.D. focuses on network coding and its applications. His research interests include network coding, computer networks, wireless sensor networks, and network security. E-mail: alfruwitonze@gmail.com

Aloys Fashaho holds a Ph.D. degree in soil science from Egerton University (2020). He holds a master’s degree in sanitary engineering/agricultural sciences and biological engineering from the “Faculté Universitaire des Sciences Agronomiques de Gembloux/Belgium” (2008) and a bachelor’s degree in soils and agricultural engineering from the former National University of Rwanda (NUR) (2002). Aloys is a lecturer at University of Rwanda, College of Agriculture, Animal Sciences and Veterinary Medicine, School of Agriculture and Food Sciences. He is also the Head of the Department of Soil Sciences. His research interests include soil fertility and conservation and agricultural waste management. He has worked on “Evaluation of Soil Properties and Response of Maize (Zea mays L.) to Bioslurry and Mineral Fertilizers in Terraced Acrisols and Lixisols of Rwanda.” E-mail: aloysfashaho@gmail.com

Adrien Turamyenyirijuru is a researcher and lecturer at the College of Agriculture, Animal Sciences and Veterinary Medicine, University of Rwanda. He received a B.Sc. in agriculture from National University of Rwanda in 2007, an M.Sc. in sustainable soil resource management from University of Nairobi in 2013, and a Ph.D. in agronomy from Egerton University in 2020. He has worked on sustainable soil management, sustainable fertilizer use, plant nutrition, and precision agriculture. He is currently the Team Leader of the Task Force in the process of operationalization of the ACES and Coordinator of Potato STIC and Dairy STIC. E-mail: adratur2005@yahoo.fr

Article

Original Article

International Journal of Fuzzy Logic and Intelligent Systems 2023; 23(2): 214-228

Published online June 25, 2023 https://doi.org/10.5391/IJFIS.2023.23.2.214

Copyright © The Korean Institute of Intelligent Systems.

Prediction of Soil Quality in Rwanda for Ideal Cultivation of Potato () Using Fuzzy Logic and Machine Learning

Christine Musanase1,2, Anthony Vodacek2,3 , Damien Hanyurwimfura1,2 , Alfred Uwitonze1,2 , Aloys Fashaho1,2 , and Adrien Turamyemyirijuru1,2

1African Center of Excellence in Internet of Things, University of Rwanda, Kigali, Rwanda
2Chester F. Carlson Center for Imaging Science, Rochester Institute of Technology, Rochester, NY, USA
3College of Agriculture, Animal Sciences and Veterinary Medicine, University of Rwanda, Musanze, Rwanda

Correspondence to:Christine Musanase (musanasechristine@gmail.com)

Received: March 23, 2023; Revised: May 23, 2023; Accepted: June 15, 2023

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

The ability to estimate soil quality has great value for agriculture, especially for low-income regions with minimal agricultural and financial resources. This prediction provides users with information that is useful in determining whether the soil is suitable for a specific crop, such as potato (Solanum tuberosum). Farmers in Rwanda lack information on soil quality. There are not enough soil laboratories to perform the requisite measurements of NPK, pH, and organic carbon, nor are there enough experts to analyze the data and provide farmers with timely results. The prime objective of the proposed study is to develop a predictive framework that can estimate soil quality for the ideal cultivation of potato (Solanum tuberosum) considering a case study of Rwanda. In this study, bootstrapping is used to augment the small soil dataset, and fuzzy logic is used to label soil data into four classes of soil suitability, with verification of the labeling by soil experts. Several machine learning methods are then tested on the labeled data, resulting in the classification of suitability for the augmented dataset and an assessment of their performance as a way to support experts in predicting soil quality. All machine learning methods applied were viable, with the best performance achieved using an artificial neural network. The quantified outcome showed that the adoption of a neural-network-based scheme has an average accuracy of 32% in contrast to other learning schemes. However, 70%-80% accuracy was achieved upon the adoption of fuzzy logic.

Keywords: Soil quality, Fuzzy logic, Artificial intelligence, Rwanda, Machine learning, NPK, Predictive model

Fig 1.

Figure 1.

Structure of the proposed scheme of implementation.

The International Journal of Fuzzy Logic and Intelligent Systems 2023; 23: 214-228https://doi.org/10.5391/IJFIS.2023.23.2.214

Fig 2.

Figure 2.

Study area from which the soil data were derived, i.e., Rubavu, Burera, Gicumbi, and Rwamagana.

The International Journal of Fuzzy Logic and Intelligent Systems 2023; 23: 214-228https://doi.org/10.5391/IJFIS.2023.23.2.214

Fig 3.

Figure 3.

Architecture of fuzzy logic type 2.

The International Journal of Fuzzy Logic and Intelligent Systems 2023; 23: 214-228https://doi.org/10.5391/IJFIS.2023.23.2.214

Fig 4.

Figure 4.

Histogram of OC percent.

The International Journal of Fuzzy Logic and Intelligent Systems 2023; 23: 214-228https://doi.org/10.5391/IJFIS.2023.23.2.214

Fig 5.

Figure 5.

N percentage histogram.

The International Journal of Fuzzy Logic and Intelligent Systems 2023; 23: 214-228https://doi.org/10.5391/IJFIS.2023.23.2.214

Fig 6.

Figure 6.

P percentage histogram.

The International Journal of Fuzzy Logic and Intelligent Systems 2023; 23: 214-228https://doi.org/10.5391/IJFIS.2023.23.2.214

Fig 7.

Figure 7.

K percentage histogram.

The International Journal of Fuzzy Logic and Intelligent Systems 2023; 23: 214-228https://doi.org/10.5391/IJFIS.2023.23.2.214

Fig 8.

Figure 8.

Water pH histogram.

The International Journal of Fuzzy Logic and Intelligent Systems 2023; 23: 214-228https://doi.org/10.5391/IJFIS.2023.23.2.214

Fig 9.

Figure 9.

Results of the fuzzy logic labeling for the three different samples. The classes are ordered by prevalence.

The International Journal of Fuzzy Logic and Intelligent Systems 2023; 23: 214-228https://doi.org/10.5391/IJFIS.2023.23.2.214

Fig 10.

Figure 10.

Comparison of all algorithms.

The International Journal of Fuzzy Logic and Intelligent Systems 2023; 23: 214-228https://doi.org/10.5391/IJFIS.2023.23.2.214

Algorithm 1. Algorithm for labeling using fuzzy logic type-2..

Input: Soil Data (S)
Output: Soil Quality

1.Add 4 new columns in S
 -c1 for pH quality
 -c2 for N quality,
 -c3 for P quality
 -c4 for K quality
2.Assign quality values according to the study conducted
3.If pH quality is 1
 -if N quality is 1 then quality is 1
 -else quality is smallest among p & k
4.Else if pH quality is 2
 -if n quality is 1 then quality is 1
 -else quality is 2
5.Quality is equal to pH quality

Table 1. Dataset sample presentation of values obtained from soil analysis.

pHOC (%)N (%)P (ppm)K (ppm)
52.880.0918.880.3
4.892.930.0729.373.5
4.942.910.078.3745.1
5.32.780.1018.888
5.02.710.0713.651
4.942.650.0778.3787

Table 2. Classification of selected soil properties values for potato.

SuitableModerately suitableMarginally suitableNot suitable
K (ppm)>5535–5515–35<15
P (ppm)>106.5–102.5–6.5<2.5
N (%)>0.300.225–0.300.125–0.225<0.125
OC (%)>0.70.5–0.70.3–0.5<0.3
pH5.5–75–5.54–5<4
7–7.57.5–8>8

Table 3. Numerical outcome of accuracy-based comparative analysis.

ML approachesPrecision (%)Recall (%)F1-score
Gaussian NB0.820.830.84
ANN0.870.850.86
Logistic regression0.790.750.77
KNN0.750.720.73

Share this article on :

Related articles in IJFIS