Expert System for Disease Identification in Palawija Plants with the Dempster Shafer Method

ABSTRACT


INTRODUCTION
Crops are one of the important commodities in agriculture that contribute greatly to the food needs of the community.However, disease attacks on crops are one of the main factors that cause a decrease in production and yield quality.Therefore, the development of an expert system for identifying diseases in crops is very important to help farmers and agricultural experts in overcoming this problem.
In this modern era, the development of information technology has made significant contributions in various fields including agriculture.One technology that has been widely applied in agriculture is the expert system.An Expert System is a system designed to imitate or illustrate the expertise and knowledge of one or more experts, to provide alternative solutions to the problem to be solved.[1] Expert systems have been used extensively in agriculture, including in the identification of diseases in crops.Using expert systems, farmers and agricultural experts can quickly and accurately identify diseases affecting crops and provide appropriate control measures.The advantages of using this expert system include speeding up the diagnosis process, reducing human error, increasing agricultural productivity, and optimizing the use of resources [2,3,4].
Based on previous research, several approaches have been taken in the development of expert systems for disease identification in plants.Methods such as fuzzy logic, [5]Certainty Factor [6], and Forward Chaining [7] have been used in various studies to address this problem.However, when compared to the Dempster-Shafer method, these approaches still have some limitations.Dempster-Shafer is a mathematical theory for proofs based on the functions of belief and reasonable thinking, used to combine separate pieces of information (proofs) to calculate the probability of an event.In the context of disease identification in crops, uncertainty and lack of clear information often occur because the symptoms of the disease often overlap or are difficult to identify with certainty [8].
Using the Dempster-Shafer method, expert systems can combine different types of knowledge and provide a degree of confidence or confidence in any given diagnosis.This method can handle situations where there is vagueness or contradiction in the data or information provided.Therefore, the application of the Dempster-Shafer method in the expert system of disease identification in crop crops is expected to increase the accuracy and reliability of diagnosis results.
This research focuses on developing an expert system for disease identification in crop crops, using the Dempster-Shafer method.Despite previous research in this area, the application of the Dempster-Shafer method to crops is still rare.This study will collect and analyze data on disease symptoms, environmental factors, and control measures.This data will be classified that represent the types of diseases that are common in crop crops.
This expert system will be implemented in a web programming language, allowing farmers and agricultural experts to input observed symptoms.The Dempster-Shafer method will be used to calculate the diagnosis of possible diseases and provide recommendations for appropriate control measures.The goal is to help farmers and agricultural experts identify diseases more accurately and quickly take necessary action.
The novelty of the study lies in the application of the Dempster-Shafer method, which has not been widely used in this context.The results of this research are expected to increase the productivity and quality of crops, as well as open opportunities for the application of this method in other agricultural fields.

Expert System
An expert system is a computer program that simulates the behavior of humans or organizations that have expert knowledge and experience in a particular field.Usually such systems contain a knowledge base containing accumulated experience and rules for applying basic knowledge to each specific situation.Expert systems are applied to support problem-solving activities, such as: Interpretation, Diagnosis, Design, Planning, Monitoring, Debugging, Instruction and Control.Another definition of expert systems is computer-based systems that use knowledge, facts and reasoning techniques in solving problems that can usually only be solved by an expert in a particular field.Here computers are used as a means to store the knowledge of experts [9,10,11].

Dempster-Shafer method
Dempster-Shafer is a mathematical theory for proof based on belief functions and plausible reasoning, which is used to combine separate pieces of information to calculate the probability of an event.The Dempster-Shafer method can be used to overcome the uncertainty factor in expert systems that causes many possible diagnoses to be found, by combining separate pieces of information to calculate the likelihood of an event based on a given confidence value.The sequence of the Demster Shafer method framework is shown in Figure 1 [12,13].
Expert System for Disease Identification in Palawija Plants… (W.Ilham, et al)

Framework
In order for research to be more directed, easy to understand, the sequence of steps will be made systematically so that it can be used as a clear and easy guideline to solve existing problems.The sequence of steps to be made in this study can be seen in Figure 2. The scope of the problem to be studied must first be determined, because if it is not able to define and determine the scope of the problem from a study, then the best solution will not be obtained.So the step of describing the scope of this problem is the most important first step in this writing.

Analyzing Problems
With problem analysis, it is expected to be able to understand the problem whose scope or limitations have been determined.

Studying literature
Literature study is an action carried out to study scientifically and theoretically the problems that have been limited previously sourced from journal literature related to research that can be accounted for.

Collecting Data
Collecting data is done by collecting all the data needed in research.The technique carried out for data collection in this study is by observation data calculation techniques.

Analyzing Data Using the Dempster-Shafer Method
To solve this problem use the Dempster-Shafer method in hopes of obtaining the appropriate similarity value based on the symptoms selected by the user.

Testing Results
Furthermore, testing was carried out on the developed data.Testing is carried out by applying the CBR Method based on existing criteria data.So that the level of system accuracy can be known by manual calculations.

Drawing Conclusions
Is the final stage of research, drawing conclusions based on literature studies and discussion of problems and is the result of analysis of research.The conclusion obtained is about the type of crop disease.

RESULTS AND DISCUSSION
In research on the identification of crop diseases using the Dempster-Shafer method which is a reasoning model that combines problem solving, understanding and learning as well as combining overall with memory processing [17].The flowchart of the Dempster-Shafer process in this study is shown in Figure 1.

Research Data
The research data used are in the form of diseases of crops used as research objects, types of crops, diseases in each crop and also symptoms of each disease.To describe the completion of the methods used in this study and the knowledge gained from interviews with experts.So information is obtained in the form of knowledge and rules for the system to make a decision based on existing categories as in Table 1.Pests and Diseases.The processed data is a disease of corn crops.

TP01: Corn
In this study, for the search process with the Dempster-Shafer method displayed symptoms on ornamental plants, especially on orchid plants which will be shown in Table 2.In the Expert System the Inference Engine is the rule for fact matching.The inference engine starts its tracking by matching the facts in the knowledge base with rules that have been defined based on the Knowledge Base.

Symptom Input
The stages of selecting symptoms that appear in orchids, each is given weight based on the user's answer.If the answer is NO then weight = 0, if the answer is YES then the weight given depending on the weight of each symptom can be seen in Table 2.The formation of rules is a rule used by an expert system obtained from experts, where the experts in question are experts on pests and diseases of orchids.Here is the form of the rule obtained: After obtaining the rule from the expert, then the rule execution process is carried out based on the answer to the user's selected symptoms.The rule execution process is carried out by entering the symptom values of the user into the rule that has been set, then from these symptoms the minimum value of all symptoms that meet the rule is found then multiplied by the value of each rule.a. Execution Rule 1 for TP01 disease IF G01 AND G02 AND G03 AND G04 AND G05 THEN P01 G01 (YES = 0.22), then : M1{P1} = 0.22; M1{} = 1 -0.22 = 0.78.G03 (YES = 0.17), then : M2{P1} = 0.17; M2{} = 1 -0.17 = 0.83.From the calculation above, the density value of risk of leaf spot is 0.18 or 18.00%.
d.In rule 4 the execution of the rule is not done because all user answers are NO.
f.In rule 6 the execution of the rule is not done because all user answers are NO.
g.In rule 7 the execution of the rule is not done because all user answers are NO.
h.In rule 8 the execution of the rule is not done because all user answers are NO.
In symptoms G38, G39, G40 and G41 do not need to be processed because the answer is No. From this calculation, the value of the risk density of Nematodes is 0.17 or 17%.
From the results of the Dempster-Shafer Process, it was found that, Corn Plants can be identified as affected by disease and the risk of late blight, which is with a value of 0.6022 or 60.22%.This expert system diagnostic testing is carried out with randomized sample trials applied to test the appropriateness of symptoms and actual diagnosis.This test takes examples of orchid plants that have different types and symptoms of pests and diseases.The author tested 10 samples to conduct trials, after the trials produced 8 samples were properly exposed to pests and diseases according to the diagnosis which means the accuracy rate reached 80%.The results of this expert system diagnostic testing evaluation are shown in Table 4.
Table 4. Test Results

Figure 1 .
Figure 1.Working Steps of the Dempster-Shafer Method

Figure 2 .
Figure 2. Framework The explanation of the research flow above is as follows: 1. Describe the Scope of the ProblemThe scope of the problem to be studied must first be determined, because if it is not able to define and determine the scope of the problem from a study, then the best solution will not be obtained.So the step of describing the scope of this problem is the most important first step in this writing.2.Analyzing ProblemsWith problem analysis, it is expected to be able to understand the problem whose scope or limitations have been determined.3.Studying literature

Table 1 .
Pests and Diseases of Palawija Plants

Table 2 .
Symptoms of Pests and Diseases of Corn Plants

Table 3 .
User Consulting