Implementation of Data Mining Using C4.5 Algorithm for Predicting Customer Loyalty of PT. Pegadaian (Persero) Pati Area Office

Abstract

PT Pegadaian (Persero) is engaged in the business of providing credit services with pawn, non-pawning and gold investment products. One of the right marketing strategies to survive today's high competition is to maintain customer loyalty, researchers use several data variables available in the MIS (Management Information System) in the form of customer transaction frequency, how many products are taken by customers, customer satisfaction and direct interviews. to predict customer loyalty of PT Pegadaian (Persero) by implementing the c4.5 algorithm. The c4.5 algorithm is the algorithm used to create a decision tree. Decision trees are a very powerful and well-known method of classification and prediction. The decision tree method converts very large facts into a decision tree that represents the rule. Rules can be easily understood in natural language. This study aims to determine the accuracy of the C4.5 algorithm to predict customer loyalty of PT Pegadaian (Persero) and the most influential factors in loyalty. The results of the experimental application of the c4.5 algorithm show that the level of accuracy generated in predicting customer loyalty is quite high, namely 89.94% in data testing 1 and 94% in data testing 2. The application of the c4.5 algorithm in predicting customer loyalty of PT Pegadaian (Persero) can be well applied.

Performance Analysis of Tree-Based Algorithms in Predicting Employee Attrition

“…The application of the C45 algorithm can be optimized, such as the addition of swarm particle optimization which achieves an accuracy of 97.13% in measuring the readiness of junior high school students to face the national exam [25] . Another result of the C45 algorithm in predicting customer or customer loyalty is reflected in 2 similar results, namely the C.45 algorithm has good accuracy [26] compared to other algorithms such as Naïve Bayes [27] , although in certain cases the results are sometimes contradictory, the resulting accuracy is still relatively good [28] . Apart from the C45 and Random Forest algorithms, the LMT algorithm is one of the algorithms that can be used for the classification process, such as in Natuthe ral Language Processing-based Mental Health Risk Prediction study, the most accurate prediction results were achieved in the DASA dataset using the sentiment dictionary and the LMT and SVM algorithms [29] .…”

Analysis of Land Plot Sales Using the C4.5 Algorithm in Property and Housing Sector

“…The attribute with the highest gain value is selected as the root of the decision tree (Rohman & Rufiyanto, 2019). In recent research, the use of the C4.5 algorithm has proven effective in a variety of applications, including in sales analysis and consumer behavior prediction (Fadhila & Hasugian, 2022;Muttaqien et al, 2021).…”

Klasifikasi Penyakit Hiperkolesterol Menggunakan Algoritma Decision Tree C4.5

“…Pada tahap ini algoritma C4.5 mempunyai 2 aturan kerja yaitu: Pembuatan pohon keputusan, dan pembuatan aturan (rule model). Aturan yang dibentuk dari pohon keputusan akan membentuk suatu kondisi berupa if maka [2]. Algortima C4.5 mengolah data dari pohon keputusan menjadi aturan sederhana yang mudah dipahami.…”