Implementation of Certainty Factor in an Expert System for Diagnosing Pests and Diseases of Tomato Plants

Tomatoes are a plant that is currently widely planted by farmers. Environmental factors and stable selling power are the reasons why this plant is popular with Indonesian farmers. However, that doesn't mean tomato plants don't have growth problems. Diseases such as leaf rot, fusarium wilt, bacterial wilt and leaf scorch are still the main factors inhibiting the decline in yield and quality of tomato plants in Indonesia. This is because farmers do not understand the correct diagnosis of tomato diseases. Misdiagnosis of diseases causes the use of pesticides that are not appropriate, resulting in damage or failure of tomato plants. When diagnosing tomato diseases, you need a farm advisor who can accurately diagnose tomato diseases. In this study, an expert system for diagnosing tomato plant diseases was built to determine pest and disease diagnoses and provide solutions and suggestions for existing diseases based on the selected symptoms. The method used in this expert system is certainty factor. This method was chosen because the certainty factor measures the value of a hypothesis's belief in a fact. These values are divided into two parts, namely MB and MD. The results of applying the confidence factor method to an expert system for diagnosing tomato plant diseases with examples of cases of late blight diagnosis by selecting the appropriate symptoms obtained a percentage of 97%, so it can be interpreted that the use of this method has the opportunity to solve problems in tomato plants


Introduction
Tomato plants are one of the agricultural commodities which are very beneficial for the body, because they contain vitamins and minerals which are necessary for growth and health. Tomato plants also contain substances that build body tissue and substances that can increase energy for movement and thinking, namely carbohydrates, protein, fat and calories [ 1].
One of the obstacles in increasing tomato production includes fruit caterpillars, earthworms, green aphids, whiteflies, fruitflies, leaf rot, fruit rot, bacterial wilt, fusarium wilt and brown spots on the leaves. This becomes a problem if it is not handled properly because it can cause the plants to die or not grow well and ultimately result in crop failure. Beginner farmers and lay people often ignore this because their lack of knowledge regarding this matter usually occurs during the planting period, and it is difficult to consult with experts in the field of plant diseases due to the limited time an expert or expert has to provide counseling to farmers, so farmers do not know how to control it which causes a decrease in the productivity of tomato plants [ 2]. Expert systems can be an alternative in solving these problems by translating the expertise of an expert into a system [ 3] [4].
An expert system is a computer application that is intended to assist decision making or solving problems in a specific field. This system works by using knowledge and analysis methods that have been defined in advance by experts in accordance with their areas of expertise. This system is called an expert system because its function and role are the same as an expert who must have knowledge and experience in solving problems [ 5].
An expert system is a computer program that represents and uses the skills and knowledge of one or more human experts to provide high-quality performance in a specific domain. Expert systems offer a number of benefits when compared to human experts [ 6]. An expert system is a system that uses human knowledge where this knowledge is entered into a computer and then used to solve problems that usually require human expertise or expertise [ 7].
The method used in the expert system for diagnosing pests and diseases in tomato plants is the certainty factor method. This method defines a measure of certainty regarding facts or rules to describe an expert's confidence in the problem at hand. By applying the Certainty Factor method in the diagnosis system for tomato plants, it is possible to determine the level of accuracy of the Certainty Factor method in overcoming uncertainty in the diagnosis of a disease in tomato plants [ 8].
Certainty Factor (CF) is a value used to measure expert confidence [9]. Shortliffe Buchanan was the person 2 who introduced Certainty Factor (CF) in creating the MYCIN expert system to show the magnitude of the confidence value [10]. The Certainty Factor method has the advantage that this method is very suitable for use in expert systems because the accuracy of the data being processed can be maintained because in one accuracy calculation process you can only process two data [11]. The Certainty Factor method can be an answer to expert confidence regarding uncertainty in translating information or knowledge through analysis that forms a matrix [12]. With the existence of an expert system using the Certainty Factor method, experts, patients and the general public can easily knowing the types of mental disorders suffered by people with mental disorders [12]. With this expert system, ordinary people can solve problems like problems solved by experts

Research methodology
The steps in solving the problem to be discussed can be clearly arranged, so a framework is needed. The research framework contained in Figure 1 follows: Research stages are steps that must be taken to make conducting research easier. The stages of this research are as follows:

Preliminary Research
Preliminary research can provide initial evidence that the problem we will examine in the field really exists. Therefore it takes time for data collection, research time, research location, research methods, field research , library research and laboratory research

Data collection
Manuscripts are written on A4 paper size with a minimum number of pages of 6 pages, a maximum of 15 pages, including tables and figures, and with reference to the writing procedures as compiled in this paper.

Problem analysis
This analysis is carried out to limit the object to be studied so that it becomes information that is more systematic and easy to understand and obtains accurate facts and will be used in research. These data will then be analyzed by the system according to what has been applied so that the system can provide the right decision.

System planning
The design is carried out using UML (Unified Modeling Language) as a tool in explaining the flow of analysis that will be made in conducting research. UML (Unified Modeling Language) which will be used use case diagrams, activity diagrams, class diagrams and sequence diagrams.

Implementation
System implementation is the stage of putting the system in place so that it is ready to operate. Implementation aims to confirm the design modules, so that users can provide input for the development of the expert system. At this stage the application design is carried out using the PHP and MySQL programming languages.

Testing
This test is focused on the functionality of the Decision Support System application which includes malfunctions, interfaces , and databases. Testing is carried out directly using the Google Chrome Web Browser program and the Mowes Portable Web Server program so that you can find out whether the results are in accordance with the expected results or not. In this trial phase it is carried out using the localhost server which is a virtual server for testing PHP-based programs.

Problem analysis
System analysis aims to limit the objects and subjects to be studied in order to become a system that is more systematic and easy to understand. To obtain data or information in this case, the authors first carry out the necessary data collection activities as a support for determining the object of research.

New System Analysis
The service process so far is said to be not good because the process of diagnosing pests and diseases is still done manually, by asking farmers directly to plant experts in Nagari Koto Laweh. Sometimes farmers can make a mistake in diagnosing plant diseases, because human memory has a limited capacity to remember all types of diseases, and farmers are still unfamiliar with tomato plant diseases. Based on the system analysis that has been described, it can be seen that it is necessary to create an expert system as a tool for diagnosing diseases in plants based on the symptoms shown by plants effectively and efficiently.

Analysis of Symptoms, Pests and Diseases
This expert system is only used to diagnose pests and diseases in tomato plants. The sample data as initial data for the types of pests and diseases detected are 8 types of pests and diseases that are often found in tomato plants. Below will be explained about the types of pests and diseases, the symptoms of the disease.

Certainty Factor Calculation Method
The first stage is to group the appropriate symptom data for a disease along with a range of MB and MD values, for example the symptoms found in late blight are as follows:   After getting the CF value from experts and users, the next step is to carry out calculations for each existing symptom, with the following results : The second calculation is obtained from the Cflama1 value which will be used as the old cf in the next calculation with the third symptom:

Consultation Page View
After registering and having an account, the user then conducts a consultation by selecting the condition of the symptoms according to the symptoms being experienced by the tomato plant. It is through this consultation page that symptoms are used as input for further processing and produce results for diagnosing pests and diseases of tomato plants.

View Consultation History
This page is a display of the user's consultation history data page.

Conclusion
An expert system for diagnosing pests and diseases for tomato plants was successfully built using the certainty factor method which was then tested for one type of disease, successfully obtaining a diagnosis result with a value of 97%. This value proves that the application of the certainty factor method is suitable and very possible to be applied in solving farmers' problems in the field. The evaluation results show that this expert system can work effectively and has a good level of accuracy in determining pests and diseases of tomato plants. As well as having the ability to identify pests and diseases based on the symptoms found on plants and provide recommendations for appropriate solutions based on the certainty factor of each symptom.