Marlinda, Recha Sri (2023) Penerapan Data Mining untuk Memprediksi Kebutuhan Bahan Bakar Minyak di SPBU Menggunakan Metode Regresi Linear Berganda. Other thesis, Universitas Islam Riau.

Text 193510331.pdf - Submitted Version Restricted to Registered users only Download (5MB) | Request a copy

Abstract

Petroleum is one of the natural resources that is indispensable to society and has a major influence on the country's economy. Petroleum processing will produce fuel oil (BBM) which is mostly used as fuel for motorized vehicles, industrial raw materials and so on. The types of fuel that are most often encountered at Public Fuel Filling Stations (SPBU) are pertalite, biosolar, pertamax turbo and dexlite. Given the increasing use of motorized vehicles, the demand for fuel is also getting higher. It often happens that fuel is insufficient so that drivers who have queued do not get fuel. To meet these needs, a system is needed that can predict how much fuel needs to be prepared. Fuel stock that matches the needs will also help sales in order to maximize revenue. In addition, a prediction system is also needed to overcome stock vacancies and minimize the occurrence of excess fuel stock. And with this system it is expected to increase revenue from pertalite sales at gas stations. Therefore, it is necessary to build an application to determine the prediction of fuel oil needs at gas stations. With this system, the results of the application of data mining for predicting fuel oil needs using multiple linear regression methods to gain knowledge in the form of regression models. Testing the accuracy of the system built has very good performance with 278 training data on each type of fuel with 70 testing data on each type of fuel obtained an accuracy of 89.02% on pertalite, 95.41% on pertamax turbo, 86.45% on biodiesel and 77.36% on dexlite so that the prediction of fuel oil needs at gas stations is feasible to implement.

Actions (login required)

View Item