CLASSIFICATION OF RESEARCH PROPOSAL FUNDING USING NAÏVE BAYES AND DECISION TREE METHODS

: In the selection process, determining a university funding research proposal at Tunas Pembangunan University (UTP) still has not fully used information technology to support related institutions, namely the UTP Institute for Research and Community Service (LPPM). So it has obstacles and requires a long time. So we need a system that is able to help these institutions to make it easier to determine recipients of research proposals that are worthy of funding. The application of data mining is a series of processes to explore added value in the form of knowledge that has not been known manually from a data set. This research has parameters, namely, NIDN, academic degree, track record, a proposed budget plan (RAB), and targeted outcomes. This is certainly less efficient because if a lecturer proposes a proposal, he must wait a long time to find out whether the results are accepted, accepted with improvements, or not. In addition, the assessment process has not used relevant methods so the results of the assessment of research proposal selection are not objective because the results of the assessment of the proposals obtained by the lecturer proposing the proposal are the final results in the form of a feasibility recommendation contained in a decision letter so that the application of classification with criteria in accordance with the selection needs is necessary. research proposal. By applying the data mining algorithm of the Naïve Bayes Method and the Decision Tree, it is


INTRODUCTION
In carrying out the realization of the Tri  algorithm with a simple formula and is easy to apply as presented by (Jadhav et al., 2016) and (Maryamah et al., 2016), while C.45 algorithm in several studies using decision tree classification, such as research (Saxena & Sharma, 2016) provides a high level of accuracy. Another study that examines the comparison of the performance of several data mining classification methods has previously been carried out (Santra & Jayasudha, 2012), in this study, used the Naive Bayes algorithm for the classification technique, whereas previous studies used the C4.5 algorithm.
Another study (Dimitoglou et al., 2012) tested the ability of data mining and  (Ginting et al., 2014). Another definition of Data Mining is a series of processes that employ one or more computer learning techniques to analyze and extract knowledge automatically or a series of processes to explore added value from a data set in the form of knowledge that has not been known manually (Sijabat, 2015).
Data mining is a term used to describe the discovery of knowledge in databases (Kusrini & Taufiq, 2009

Comparison of Data Mining Model C4
Classification Algorithm. 5 and Naive Bayes for Diabetes Disease Prediction (Fatmawati, 2016).

Comparison of C4 Algorithm
Performance. 5 and Naive Bayes for the

Classification of posts Twitter traffic
jam the city of Jakarta using algorithm C4. 5 (Hajrahnur et al., 2018).

MATERIALS AND METHODS)
This study, uses a methodological step consisting of several stages as follow: Picture The step taken at this stage is to translate the design that has been formed into a system that applies the Naive Bayes algorithm. (Jadhav et al., 2016) stated that the Naïve Bayes Classifier is an independent model that discusses simple classification based on the Bayes theorem. Naïve Bayes is an algorithm that can classify a certain variable using probability and statistical methods. Broadly speaking, the Naïve Bayes algorithm can be explained as follows:

Implementasi Decision Tree
The concept of a decision tree or decision tree is to convert data into decision rules. The main benefit of using a decision tree is its ability to break down complex decision-making processes into simple ones so that decision-making will make it easier to solve a problem (Rismayanti, 2018). Decision Tree is one of the most popular classification methods because it is easily interpreted by humans (Wahyuningsih & Utari, 2018). A Decision Tree is used for pattern recognition and is included in statistical pattern recognition (Rosandy, 2016). The Decision Tree uses 2 calculations, the first is the Gain calculation in Equation 2 and the Entropy calculation in Equation 3 (Nugraha et al., 2016).

Bayes and Decision Tree
Testing an algorithm requires standards and test equipment (Kurniawan & Kurniawan, 2018). Comparing 2 algorithms must have the same standard so that the best algorithm can be known from the comparison. At this stage, testing is carried out by calculating the value of precision, recall, and accuracy from Naive Bayes and Decision Tree. The initial step in this stage is to divide the data in each case into 2, namely training data or training data and testing data or test data. Training data is used as reference data in the calculation of each algorithm, while testing data is used to assess the predictions and determinations made by each algorithm are correct or not. In dividing the data into training data and testing data, several comparisons were made.

Algorithm Comparison
At this stage, a comparison of the values of precision, recall, and accuracy is carried out for each algorithm in each case.
After that, the results of each algorithm are recapitulated so that conclusions can be drawn regarding the best algorithm for each case.