The development of the cosmetics industry experienced a very significant development. Along with the development of cosmetics, information technology has become one of the inputs for the company. Information about the sale of cosmetics and information on cosmetic supplies is an inseparable thing. Cosmetics are products made from plants, minerals and animals. Predictions about future sales can control the amount of available product stock, so that shortages and excess product stock can be minimized. Accurate Prediction results will facilitate the fulfillment of market demand and can provide timely predictions. This research was conducted to predict cosmetic sales. Knowing the supply of cosmetics is something that can be found in cosmetic stores, to facilitate the prediction of cosmetic inventories in the coming period can use sales data in the previous period. Date time series is one of the time series methods used for forecasting methods to find out the pattern of patterns in the past to find out patterns in the future. The Support Vector Machine algorithm will work accurately when the required features are relevant. But inreality the features needed are not relevant, feature selection can be done to select the features needed to obtain accurate information, so as to increase efficiency. In this study the Support Vector Machine algorithm produces the best method, which is 0.136 based on the smallest error value with period 4 variables, Validation Shuffled Sampling 10 and Polynomial kernel type. The best model is also obtained with the Support Vector Machine algorithm with forward selection resulting in the smallest error value of 0.144 with period 8 variables, as well as the Validation Shuffled Sampling 10 and the polynomial kernel type. Keywords: Kosmetik, Data Mining,  Support Vector Machine, feature selection, forward selection

Published by Universitas Janabadra
Journal Name Informasi Interaktif
Contact Phone-
Contact Name -
Contact Email -
Location Kota yogyakarta, Daerah istimewa yogyakarta INDONESIA
Website informasiinteraktif| http://e-journal.janabadra.ac.id/index.php/informasiinteraktif|
ISSN ISSN : 25275240, EISSN : -, DOI : -,
Core Subject Science,
Meta Subject Computer Science & IT,
Meta DescJurnal Informasi Interaktif mempublikasikan artikel dalam bidang teknologi informasi dan komunikasi, rekayasa perangkat lunak dan sistem informasi.
PenulisApriliyani, Aflahah , Utami, Ema , Raharjo, Suwanto
Publisher ArticleUniversitas Janabadra
Subtitle Article Informasi Interaktif Vol 4, No 2 (2019): Jurnal Informasi Interaktif
Scholar Googlehttp://scholar.google.com/scholar?q=%2Bintitle%3A&…
View Articlehttp://e-journal.janabadra.ac.…
DOIhttp://download.garuda.ristekdikti.…
DOI Number Full PDF (343.961 KB)
Download Article [1] http://e-journal.janabadra.ac.id/index.p…
Download Article [2] http://download.garuda.ristekdikti.go.id…

 

Prediksi (acara televisi) The Prediksi Unsur-unsur prediksi Mendeleev Prediksi Kiamat 2000 Prediksi Kiamat 1999 Prediksi Kiamat 2012 Prediksi Kiamat 2008-2009 Prediksi struktur protein Prediksi Kiamat 2011 Daftar tanggal yang diprediksi sebagai hari kiamat Pemberitahuan tentang pengkhianatan terhadap Yesus Pilihan rata-rata yang diprediksi Prakiraan cuaca Prediksi pandemi dan persiapan sebelum pandemi COVID-19