Prediksi Penjualan Produk Pepsodent Unilever dengan Algoritma K-Nearest Neighbor

Abstract

In the era of globalisation and increasingly fierce market competition, companies are striving to improve their operational efficiency and marketing strategies to maintain market share and increase revenue. PT Unilever Tbk, as one of the multinational companies that operates various types of consumer products, including dental care products such as Pepsodent, requires reliable sales prediction to maximise its product performance in the market. The main objectives of this research are to apply the K-Nearest Neighbor method to Unilever pepsodent products in a prediction model that can preprocess pepsodent product data for the last 1 year using Rapid Miner and to measure the accuracy of Pepsodent product sales predictions. The data used is the number of stocks, types of pepsodent, sales, seasonal factors. From the results of analysis and evaluation, it can be concluded that the prediction accuracy in the K-NN algorithm is able to provide fairly accurate sales predictions for Pepsodent Whitening products with a value of 161, 186, 165 equally 114. Pepsodent Economy with a value of 982 predictions 1021, a value of 638 and 774 predictions are both 927. Pepsodent Herbal with a value of 173 predicted 193 and a value of 129 and 118 predicted values are both 207. Accurate sales predictions are helpful in production planning and marketing strategies, which in turn can improve operational efficiency and customer satisfaction. The K-NN algorithm proved to be effective in this case, although proper selection of the K parameter is essential to obtain the best results.