CLASSIFICATION C4.5 METHOD IN SELECTION OF PASKIBRA MEMBERS

: Paskibra selection is an annual agenda held by schools, district disporas, provincial disporas, and the Ministry of Youth and Youth. This activity aim select the best students who will serve as flag raisers at independence commemoration ceremonies and other holidays. Paskibra participants have to choose. This reliable, formidable and attractive students are choosen. So the flag raising can be carried out solemnly and well. This school has a large number of students. Paskibra extracurricular enthusiasts also have many enthusiasts. This school Paskibra is a favourite extracurricular and quite prestigious among students. The selection of Paskibra has been running so well, but the decisions taken are sometimes not by the standards procedures. Therefore, a system that will be a reference for objective decision-making need to created . The decision-making system using the data mining algorithm C4.5. This study aim to create a system that makes it easier for the selection committee to select eligible Paskibra members. The method used is a qualitative method in the form of the results of interviews with data from participants in the last three years of the Pakskibra selection. As a result, there are three main criteria and five rules of the selection system.


INTRODUCTION
One of the good impacts of the presence of information technology in the community is its ability to take over human work. Technology can make it easier for humans to do anything with the help of technology because almost all aspects of life already use technology, especially information technology. In other words, information and communication technology can facilitate and enhance the quality of human life.
Decision making implements information technology.
Decisionmaking that is previously still done manually and tended to use feelings rather than calculating, can now be taken over by computers, especially information technology. Many methods can be used in decision-making, one of which is using C4.5 classification data mining. The C4.5 algorithm generates a decision tree form . Decision trees are used in the method of classification and prediction [1]. Decision trees utilized to expose data and find hidden relationships between several prospective input variables and a target variable.
Every year the school, Dispora Kabupaten, Dispora Provinsi dan Kemenpora hold a selection paskibra. These agenda aim to select the best students who will serve as flag raisers at in-dependence commemoration ceremonies and other holidays[2]. Paskibra participants need to be choosen so that reliable, tough and attractive students are selected so that the flag raising can be carried out solemnly and well. In the beginning of each new school year, the school will begin to select students who will become members of paskibra, including SMAN 1 Kisaran. The selection of paskibra that has been running so far has been done well, but the decisions taken are sometimes not in accordance with the standards that have been set. And it is necessary to create a system that will be a reference for objective decision making.
The Data mining methods are utilized to decide covered data that is valuable to healthcare specialists with effective analytic decision making [2].
The meaning of Data Mining has the overall objective of extracting information (with intelligent methods) from data sets and converting information into understanda-ble structures for further use[3]. The C4.5 algorithm can support new student admission decision-making through the rules generated. The testing process with RapidMiner resulted in an accuracy of 90.50% [4].
An application built using the K-Nearest Neighbor (KNN) and Simple Additive Weighting (SAW) methods can recommend the names of participants who pass and do not pass the selection based on the results of ranking the scores of each participant [5]. The K-Nearest Neighbor method was used to classify chosen participants. The Simple Additive Weighting method was used to perform ranking. This application built using PHP programming language.
Determining the class of prospective students at web-based English course institutions also implements the C4.5 decision tree algorithm.
The results of the classification of students as much as 45 data played by an active class are 8, and ultimate are 6 data. Then produced a basic are 15 data, prov are 5 data. [6] Improving students' English skills also applies Algorithm C4.5. From this study it was found "hearing from the environment" as an influential factor in improving students' English skills". [7] The criteria assessed in the selection of paskibra members at SMAN 1 Range are physical strength, height, agility, discipline and marching ability. The study aims to conduct a decision support system for the selection of paskibra members using C4.5 classification data mining. The C4.5 classification data mining process data on student assessment criteria for prospective paskibra members is described, resulting in a final grade that can be used as a basis for decisions for the implementing committee.

C4.5 Method
Many algorithms that can be used formation of decision trees include ID3, CART, and C4.5. Algo-rhythm C4.5 is a development of ID3. Because of this development, the C4.5 algorithm has the same basic work principles as ID3 algorhythms. It's just that in the C4.5 algorithm, the selection of attributes using the Gain Ratio. Resulting in the Decision Tree C4.5 algorithm has the good accuracy value [8] [9].
Decision Tree C4.5 C4.5 method is to change the tree generated in several rules. The number of rules is equal to the number of paths that might be built from the root to the leaf node [6]. In general, C4.5 algorithm to build a decision tree with the following general steps: a) Select the attribute as the root b) Create a branch for each value c) Divide cases in branches d) Repeat the process for each branch until all cases in the branch have the same class [10]. To select attributes as roots is based on the highest Gain score of the available attributes [11]. To count the Gain, the formula as seen in Formula 1.

Data Analysis
This study involved SMA Negeri 1 Kisaran as a system user to determine the graduation of students in paskibra selection with criteria including height, physical strength, agility, discipline and the ability to line up with each weight or value against the percentage of excellence possessed by these criteria. In this case, result of data mining algorithm C4.5 is implemented in the decision support system .
In this study, the data collected are a. Student data of selection participants; b. Data recapitulation of the assessment of selection participants from 2020. Some of the criteria needed in determining the passing selection of paskibra participants can be seen in table 1.
In this section, the research design is no longer contained but is focused on the result of the research that has been carried out.
The result of the study must be explained clearly and concisely. The result should summarize the findings The discussion must explore the significance of the research result. It is best to quote from previous research that can support the results of your research [12].   After the entrophy value is count, the next step is count the gain value using the following equation: Gain value of Heigh : The gain physical strength, agility, discipline and marching are count as the height calculation.
. Based on the calculation of the first iteration described earlier, the obtained decision tree is in Image 1: The calculation continues to the next iteration again and again, until there is no more criteria can be used as the next branch, and all attributes in the last branch, namely discipline, have classified cases into one. From the calculations carried out produce a decision tree as in Image 2. Image 3. Data Import Wizard

RESULT AND DISCUSSION
Then this system is imported into a tool that has been determined by researchers, namely using rapidminer software. The import data in question are as seen in Image 3. In Image 3, the condition attributes of the selection of paskibra members, namely height, physical strength, agility, discipline and marching displayed. The decision attribute is a result marked by a label mark on the wizard form. After that, in the application of rapidminer software, the connection process will be carried out between the imported data and the decision tree method operator in the operator menu. Then drag the data and operators or move the process into the process window. So you can see the connection in Image 4.
From the stages are in Image 4, the application procces is running by clicking the run button in the application window. So from the results of the process run, the results are decision trees that describe the relationship between criteria that cause the passing or not of the paskibra selection participants.
Image 4. Import Data Connection with Method C4.5 Image 5. Decision Tree Results with Method C4.5 From the picture above, we get some rules as IF height does not more than165 cm THEN decision does not pass; IF height more than165 cm AND good marching ability THEN decision passes; IF height more than165 cm AND poor marching ability THEN decision does not pass; IF height more than165 cm AND medium marching ability AND discipline THEN decision passes 80%; IF height more than165 cm AND medium marching ability AND undisciplined THEN decision does not pass 80%.