JOURNAL OF SOFTWARE ENGINEERING & INTELLIGENT SYSTEMS
ISSN 2518-8739
31 December 2019, Volume 4, Issue 3, JSEIS, CAOMEI Copyright © 2016-2019
www.jseis.org
Corresponding Author: Jonah Joshua 132
Email Address: moosayimahdis@gmail.com
DIAGNOSIS MODEL OF SOY BEANS DISEASES
USING NEURO-FUZZY SYSTEM
IBRAHIM RAHMON
1
, OLUTAYO AJAYI
2
, MONDAY EZE
3
, JONAH JOSHUA
4
1,3,4
Babcock University, Nigeria
2
Federal University of Agriculture Abeokuta, Nigeria
Email: rahmon.ibrahim@yahoo.com
1
, olutayoajayi@gmail.com
2
, ezem@babcock.edu.ng
3
,
joshuaj@babcock.edu.ng
4
ABSTRACT
Soybean is an important legume crop, extensively cultivated for food on which
low-income population highly depend on its proteineous nutrient on daily basis
for food. and oil. Soy beans consumption have been the major cheap protein-rich
grain useful for treatment of malnutrition among children, for fighting against
diabetes, high blood pressure, etc. Despite the nutritional and economic value of
soy bean crop, a variety of pest attack such as fungi, nematode, bacteria and
viruses are speedily becoming a constraint to quality and bountiful harvest. The
effort of farmers to specifically identify the specific pest responsible for
damaging of plants’ segment such as roots, stem, pod and leaves still remain
vague and imprecise to many farmers. In this work, a neuro-fuzzy system was
built with MATLAB version 8 with100 rules on six input parameter as linguistic
variable or symptoms into the system to determine the disease type either as fungi
or bacteria or virus and to also determine intensity rate, that is, level of damage,
as the output in form of a crisp. The proposed Neuro-Fuzzy system was developed through MATLAB software
using Adaptive Neuro-Fuzzy Inference System (ANFIS) box. ANFIS hybridizes the learning capacity of neural
network with if-then rules of fuzzy logic to learn and design the most fitted membership function for a given set
of data and thereby map the input with output. The proposed Neuro-Fuzzy System consisted of five stage: input
stage, fuzzification, rule base, inference engine and defuzzification. The output of the system was to produce
results for the decision maker to provide solution regarding the treatment of the infected plant for bountiful and
quality harvest.
Keywords: neuro-fuzzy system; crisp; matlab; fuzzification; de-fuzzification;
1. INTRODUCTION
The technological evolution in computing has been the principal tool in increasing agricultural products.
Nevertheless, there are numerous problems and constraints working against the bountiful and quality harvest of
soybean commercial production [3]. ICTs play vital role in facilitating agricultural growth. The scientific and
technological developments, which include e-agriculture, decision support system for farmers and mobile
applications, have tremendously delivered relevant services for farmers in tackling all forms of crop diseases.
ICTs have promoted new farming techniques and distributed new knowledge using computing technology for
facilitating diagnosis and treatment of crop diseases [10].
Since the discovery of artificial intelligence (AI) theories and techniques over a decade, there have been
tremendous growths in the development of expert systems in providing solution to some uncertain and imprecise
tasks. The capacity and efficiency of expert system in imitating human reasoning process and providing relevant
advice similar to human expertise has singled it out as one of the artificial intelligence branches widely embraced
in many fields today [11]. Neuro-Fuzzy as one of the artificial intelligence methods found to be more effective in
developing medical system that provide optimal solution to problems that are vague and imprecise in nature.
Neuro-fuzzy is an efficient technique that combines the strength of two different techniques, namely; artificial
neural network (ANN) and fuzzy logic (FL) in which back-propagation algorithm of ANN performs the
computation of the fuzzy system parameters. The choice of neuro-fuzzy technique for this work is justified as a
result of need to accept six major symptoms of soybean disease as input parameters for disease classification and
computation of intensity proportion of a particular disease.
In spite of the economical value of soybeans to the farmers, nutritional advantages to the body system of
consumers, and the use of organic and inorganic fertilizers to increase total income and bountiful harvest, the main
constraints and threats challenging large production of soybean are pests attack and diseases infections.
Academic Editor: Muhammad
Jehanzeb, APCOMS,
Rawalpindi. Pakistan
Published: 31 December, 2019
Funding: Authors have no
funds.
Article Type: Research Article