Implementation of Data Mining to predict sales of Bogo helmets using the Naïve Bayes algorithm

Abstract

Consumer needs for safety and comfort in driving are very important, especially for two-wheeled or motorcycle riders. A good helmet is a helmet that is safe and comfortable when worn. Helmet qualifications that meet SNI standards are open helmets and closed helmets. Transactions are carried out online, because it was still in a pandemic situation when it was established. By carrying the tag line "Ride Safety With Your Own Style", trying to educate the younger generation to keep paying attention to safety when driving but also not neglecting fashion. For the type of sale of bogo and retro helmet brands, with a variety of colors and affordable prices. A common problem faced is how to predict or forecast future helmet sales based on pre-recorded data. This prediction is very influential on the decision to determine the number of helmets that must be provided, if you order helmets in sufficient quantities and it turns out that only a few helmet sales are sold and this will cause the stock of helmets to accumulate. The results of predictions for the sale of bogo helmets in Baris True Instances amounted to 16 data, which means Valid with 53% data accuracy. Meanwhile, there are 10 data classified as Instancely Classified Instance which means Invalid, with data accuracy of 46.67%. The amount of accuracy in the Weka application is the same as the amount of accuracy in Excel calculations. From the explanation above, the Naïve Bayes algorithm method is the best solution for predicting important things in a business need and others.