Energy evaluation and prediction system based on data mining

With the development of economy, energy using has attracted more and more attention. To help decision-makers and governors manage the energy utilization better, we write this paper to introduce Energy Evaluation and Prediction System (EEPS). Based on the data provided by SEDS, this paper proposed four models to select and aggregate important information. Firstly, Energy Profile Model (EPM) is established to cluster the data. The data is divided into for parts: production, consumption, unit price and total expenditure. Secondly, based on EPM, time is considered and we get the main energy percentage diagram of each state during 50 years. To help the governors understand the similarities and differences of the four states in using clean and renewable energy, we establish Energy Correlation Analysis Model (ECAM) to study the correlation between new energy using and the factors. Thirdly, to determine which of the four states appeared to use clean energy best in 2009, New Energy Profile Model (NEPM) is established. We suggest an objective function of different energy on production and consumption. After that we use TOPSIS to get the best solution, which shows AZS use clean energy best. Fourthly, Energy Profile Prediction Model (EPPM) is established to predict the energy profile of 2025 and 2050. We use BP and LSSVM algorithm in the model. From the prediction results, AZS will produce most of the petroleum products by 2025, and renewable energy will account for one quarter of the energy used. By 2050, the production of electricity and fossil fuels will be the main source of energy. Fifthly, EPPM and NEPM are used to predict and evaluate the use condition of energy in 2025 and 2050. From the prediction results, the clean energy used is increasing.