Expert System for Diagnosing Corn Plant Diseases Using the Web-Based Certainty Factor Method

: In the process of corn plant growth, it often faces the threat of various diseases that can hinder optimal growth. Dealing with this issue often requires the presence of a plant expert, such as a field agricultural extension officer. However, limitations in the human resources of agricultural extension officers, time constraints, and geographical distances often make it difficult for farmers to get direct assistance from experts. To overcome these constraints, a system capable of addressing issues related to corn plant diseases, similar to the role of an expert, is needed. In this study, the Certainty Factor method is employed, suitable for an expert system that measures the level of uncertainty, and through testing with 28 sample data cases, the accuracy level of this system reaches 90%.


Introduction
Rapid advances in science and technology have had a significant impact on human work, including providing a lot of assistance to the work of farmers and cultivation so that it can be done more easily.Corn, as one of the strategic cereal crops and has high economic value, also has the potential to be developed into processed products (Eldad, 2022).This plant has an important role as the main source of carbohydrates and protein after rice.Almost every part of the corn plant has economic value, both as human food, animal feed, industrial raw materials, and for food, beverages, paper, oil and bioethanol production.Apart from that, corn stalks are also a very potential source of animal feed (Kholilah, 2023).
In Indonesia currently, there are still challenges in the food sector, such as changing the function of agricultural land into industrial and residential areas which has an impact on reducing corn productivity.In addition, uncertainty in seasonal changes and farmers' limited ability to deal with pest and disease attacks can contribute to a decline in corn crop productivity.
Corn plants are at risk of being attacked by pests and diseases that can appear at any time.Some diseases that can threaten corn plants include leaf blight, sheath rot, downy mildew, cob rot, and many more.Lack of corn production can be one of the causes, and can even cause crop failures of up to 90% (Syarifuddin, 2016).The use of expert systems can provide significant assistance in dealing with pest and disease problems in corn plants.This system works by identifying the symptoms that appear, deducing the type of pest or disease that is attacking, and providing information for handling the problem (LD.Bakti, 2021).The appropriate method for an expert system in the context of corn is the Certainty Factor Method.This method uses clinical parameter values, introduced by MYCIN, to indicate the level of confidence.Certainty Factor (CF) can vary in various conditions, one of which is when there are several antecedents (in different rules) with one consequence (Imran, B. 2022).
The certainty factor method is a method to prove the uncertainty of an expert's thinking, where to accommodate this, it describes the expert's level of certainty regarding the problem at hand (Rosiana, 2023).
There are several previous studies written by other researchers that discuss expert systems that are used as reference material in this research.
Minarni, I. Warman in the journal Implementation of Case-Based Reasoning as an Inference Method in an Expert System for Identification of Corn Plant Diseases, the test results show that the system is able to identify corn plant diseases with symptoms according to the rules of 100%, and the level of accuracy with the nearest neighbor similarity method is 74 .63 % (Minarni, 2018).Then Wicaksono et al (2018) carried out disease diagnosis in soybean plants using the Dempster-shafer system testing method on 25 test cases showing an accuracy of 92%.This shows that the system is good and can be used to diagnose diseases in soybean plants (Rahmat, 2018).
In Marcelino Oktviansyah's, 2022.research entitled Expert System for Diagnosing Eye Diseases Applying Certainty Factor and Forward Chaining Methods, testing was carried out by testing 40 data samples based on existing symptoms, carried out by system and also manually to see the accuracy of the values given.The results provided have a fairly high level of certainty in the diagnosis of hypermetropia with a certainty value of 88.43% and the lowest certainty value in the diagnosis of presbyopia with a value of 61.702%.From all total tests, the probability certainty level was found to be around 51% -79% .
The difference between this research and the previous one.The title of the research that will be made is an expert system for diagnosing diseases in corn plants using the web-based certainty factor method.The Certainty Factor method is a method that defines a measure of certainty regarding facts or rules to describe an expert's confidence in the problem at hand.The advantage of the certainty factor is that it is suitable for use in expert systems that contain uncertainty, such as corn plants which are often affected by changes in weather and climate.
The aim of the current research is to help identify the types of diseases in corn plants suffered by farmers' corn plants and also provide suggestions to speed up handling of disease symptoms in corn plants and provide a health consultation facility website service that is easy for farmers to understand.And in this research an expert system will be built to identify diseases in corn plants using the web-based Dempster Shafer method.This research design used PHP, MYSQL, Bootstrap, XAMPP, UML, Flowchart, ERD tools.With the results when the user reports things that arise regarding the symptoms experienced by the corn plant and this expert system combines the rules that are already available in the database or with information, the expert system, then provides additional information such as the symptoms experienced by the corn plant, and the program output results in the form of information about disease diagnosis in corn plants and will also provide solutions for appropriate treatment.With this expert system, it is hoped that farmers can be helped to identify diseases in corn plants and provide appropriate treatment.

Research Method 2.1 Software Development Methods
In developing an expert system to diagnose corn plant diseases, the Waterfall software development model was chosen.The selection of the Waterfall Model was carried out because the process was sequential and systematic, with the following stages (Wahyudi, E. 2023) Figure 1.Waterfall Method Referring to Figure 1 above, there are 5 stages in the waterfall method, namely; 1) Requirements Analysis: In this phase, software requirements are collected and identified in detail.This involves interaction with stakeholders to understand system requirements.2) Design: Once the requirements are gathered, this phase involves designing the system architecture and detailed design of the software components (Wahyudi, E. 2023).3) Development: In this phase, the software code is actually created based on the design that has been made previously.It involves writing code, unit testing, and component integration.4) Testing: After implementation, the software is tested to ensure that it meets the requirements established in the requirements phase.Testing can include functional testing, performance testing, and so on.5) Maintenance: After implementation, the software enters the maintenance phase where bug fixes, upgrades, and other routine maintenance are required (Wahyudi, E. 2020).
The waterfall method is suitable for projects that have stable and clear requirements from the start, and where changes in requirements tend to be minimal.However, its weakness is that it is less adaptable to changes in requirements that may arise during the development process (Wahyudi, E. 2024).

Certainty Factor Method
According to (T.Sutojo, Artificial Intelligence" 2011) Certainty Factor (CF) theory is to accommodate the uncertainty of an expert's thinking (inexact reasoning) which was proposed by Shortliffe and Buchanan in 1975.An expert (for example a doctor) often analyzes existing information with expressions of uncertainty, to accommodate this we use the certainty factor (CF) to describe the level of expert confidence in the problem being faced.In expressing the degree of certainty, certainty factor assumes the degree of certainty of an expert regarding data.This concept is then formulated in the following basic formulation: CF

Inference mechanisms
The inference mechanism is one part of an expert system that uses a list of rules based on certain sequences and patterns to make inferences based on arguments.The quality of the seeds is laborious.0.8 0.6

Certainty Factor Calculation
By applying the certainty factor method, information can be found regarding diseases that are attacking corn plants.By referring to the weight values, the system will carry out calculations to determine a diagnosis that matches the symptoms entered by the user.

Results and Discussion
After determining the weight and rule values, we then go into the case with examples of 3 symptoms in each 3 diseases below: 1) There are lesions or spots on the leaves of corn plants.These lesions are initially small and brown or grayish green in color 2) Infected leaves may dry out and fall early 3) Leaves change color to yellowish or brownish 4) The fronds or leaves that surround the petiole look rotten, soft, and change color to dark brown or black 5) Bad smell around the plant 6) The infection spreads to the segments or joints between the leaf stalks and stems of the corn plant 7) Infection appears on the leaf stalks and stems of corn plants 8) There are dark dots on the surface of the lesion or around infected leaves and petioles 9) The upper and lower surfaces of the leaves are white like powder So the calculation method is as follows: 1) P01 (Leaf Blight) G01 = There are lesions or spots on the leaves of the corn plant.These lesions are initially small and brown or grayish green in color.So from the symptoms experienced, the sufferer is suffering from leaf blight and another possibility is corn cob due to leaf blight got a percentage of 95%, frond rot 90% and corn husk 94%.

System Implementation
After determining the disease, method and calculation of the certainty factor, we created an expert system on corn plants based on a website that will then be implemented in this research.The website display can be seen in picture 2 below.In this website there are various features including, home page, diagnosis page, disease register page, knowledge base page, consultation history page, disease page and disease symptom page.After the system was built, we tested the system to test whether the built system was running and could be used in accordance with what the researcher expected.This testing is done on all menus and features on each website page using the whitebox testing method (Suryadi, E. 2018).The results of system testing can be seen in table 4 below.

Diagnosis menu
Press the diagnosis menu, select symptoms and press the process button.

Print the diagnosis report
Click the print button on the diagnosis page.
Print the diagnosis results

Conclusion
From the results of applying the system using the Certainty Factor method to diagnose eye diseases in corn plants, several conclusions can be drawn.The Certainty Factor method was successfully integrated into the expert system for the diagnosis of corn plant diseases.The developed expert system is capable of diagnosing 7 types of diseases on corn plants by referring to the knowledge of three experts, and providing diagnostic results that are consistent with manual calculations.
This expert system adopts the Certainty Factor method to diagnose diseases on corn plants, which makes it possible to provide diagnosis results efficiently with a level of certainty for each disease.This system is able to diagnose diseases based on the belief value of each disease using the Certainty Factor formula.From the calculation results of the system diagnosis, it can be concluded that the accuracy level of the system reaches 90% when compared to the diagnosis made by a doctor.
[H,E] = MB[H,E] -MD[H,E] Description: CF = Certainty factor in hypothesis H which is influenced by fact E MB(H,E) = measure of belief in hypothesis H, if given evidence E (between 0 and 1) MD(H,E) = measure of disbelief in evidence H, if given evidence E (between 0 and 1) H = Hypothesis E = Evidence (event or fact) ] = user confidence measure CF(E) = measure of expert confidence And the rules for the same conclusion are: CF combination CF[H,E1] = CF[H,E1]+[H,E2] * (1-CF[H,E1]) CF combination CF[H,E]old3 = CF[H,E]old +[H,E]3 * (1-CF[H,E] old) CF[h,e] = CF(user) -CF(Pakar) = 0.7 -0.3 = 0.4 * 1 = 0.4 G02 = Infected leaves may dry out and fall early or leaves that surround the leaf stalk look rotten, mushy, and change color to dark brown or black CF[h,e] = CF(user) -CF(to the joint or joint between the leaf stalk and the stem of the corn plant CF[h,e] = CF(user) -CF(appears on the stalks and stalks of corn plants CF[h,e] = CF(user) -CF(Pakar) = 0.9 -0.1 = 0.8 * 1 = 0.8 G011=There are dark colored dots on the surface of the lesion or around the infected leaves and petioles and lower surfaces of the leaves are white like flour CF[h,e] = CF(user) -CF(

Figure 2 .
Figure 2. The diagnosis page

Table 1
Table continues on the next page..