Perbandingan Algoritma C4.5 dan Classification and Regression Tree (CART) Dalam Menyeleksi Calon Karyawan

Abstract

This research compares the accuracy of the C4.5 algorithm and Classification and Regression Tree (CART) for prospective employees selection in companies. This research using dataset with criteria like age, working experience, recent education, marital status, number of abilities possessed, and the result of admission selection test. Testing uses 200 prospective employee selection data manually from a company. Algorithm testing using K-Fold Cross Validation and the accuracy calculation of the algorithm using Confusion Matrix. C4.5 algorithm has a level of accuracy, the success rate of the system, and the level of accuracy of the decision results of 52,83%, 41,48% and 43,98%, and CART algorithm is 53,33%, 44,06%, and 42,81%.

Perbandingan Algoritma C4.5 dan Adaptive Boosting dalam Klasifikasi Penyakit Alzheimer

“…Dalam algoritma ini, setiap node menunjukkan nilai atribut, setiap cabang menunjukkan nilai atribut, dan setiap daun menunjukkan kelas. Berbagai langkah yang diambil untuk membuat pohon keputusan algoritma C4.5 adalah sebagai berikut [18]: 1. Tentukan atribut sebagai akar Setiap atribut yang dipilih untuk menjadi akar berdasarkan nilai gain ratio tertinggi dari semua atribut [19].…”

Classification and Regression Trees (CART) Algorithm for Employee Selection

“…K-Fold Cross Validation [2] is a method for evaluating the performance of a model by separating the data into two subsets, namely training data and evaluation data. In this method, each fold leaves a set of training data that will be used for the evaluation process.…”

“…The large number of applicants will make the company have many choices in recruiting new employees. With the competence of various applicants, it will make it easier for companies to choose which employee candidates are right to be recruited [1], [2]. So that in the future, it is hoped that these employees will be able to contribute well and increase the company's income financially.…”

“…There are several algorithms in classification that have been proposed by many literatures such as Decision Tree, Naïve Bayes, K-Nearest Neighbor. In the Decision Tree algorithm, there are several types of algorithms including C4.5 [5], Classification and Regression Trees (CART) [2], [6], Credal C4.5, and Adaptive Credal C4.5, and ID3 [7], [8]. Previously, research showed that the CART algorithm has a performance advantage over K-Nearest Neighbor for predicting the area of rice harvested land.…”