Penerapan Data Mining untuk Memprediksi Jumlah Produk Terlaris Menggunakan Algoritma Naive Bayes Studi Kasus (Toko Prapti)

Abstract

Toko Prapti is a small privately owned company that sells basic necessities,. So far, the prapti shop produces sales data every day, but the results obtained show that the prapti shop has not maximized the data so that it becomes a data accumulation. Therefore, the researcher conducted a study on product sales data by utilizing and applying data mining using the nave Bayes classifier algorithm to determine the interest in purchasing goods at the prapti shop. data. In this study, the author uses the waterfall system development method. The author implements this research using a web programming language, namely PHP, using the CodeIgniter framework with MySQl database. The system built with the nave Bayes algorithm includes product sales data, nave calculations of each attribute and reporting. This system produces 4 attributes that greatly affect the results of the classification. The attributes used in this research are the attributes are quarter 1, quarter 2, quarter 3 and quarter 4. Prediction results obtained using the nave Bayes algorithm produce information that can be used by stores to identify the best-selling products purchased by consumers so that it can help prapti shops to find and determine the target market more accurately. Sources of data taken from the previous 1 year with system accuracy using a confusion matrix resulted in 83.3% accuracy, 84.2% precision and 88.9% recall.   

 

Keywords : Data mining, Nave bayes Classifier, Code Igniter, Confusion Matrix