Search

This research aims to determine the influence of Advertising, Sales Promotion, Personal Selling, Public Relations, and Direct Marketing on Purchasing Decisions at the Konde Mart minimarket, sample of 100 respondents. This research uses quantitative with an Pengaruh Periklanan, Promosi Penjualan, Penjualan Personal, Hubungan Masyarakat dan Pemasaran Langsung Terhadap Keputusan Pembelian di Minimarket: Studi Kasus Konde Mart online questionnaire via Google Form. This research was conducted using multiple linear analysis methods and tested using SPSS Version 25. The test results showed that the advertising variable (X1) had a positive and insignificant effect on purchasing decisions, the Sales Promotion variable (X2) had a positive and significant effect on purchasing decisions, the Personal Sales variable (X3) had a positive and insignificant effect on purchasing decisions, variable Public Relations (X4) has a positive and significant effect on purchasing decisions, the Direct Marketing variable (X5) does not have a positive and significant effect on purchasing decisions. [ABSTRACT FROM AUTHOR]

This research paper proposes a new hybrid methodology for very short-term load forecasting of hourly or designed to predict load for 1 hour ahead. The proposed hybrid methodology is based on weather data especially for optimizing the operation of power generating electricity from thermal generation. This hybrid modelling is a combination of the Fuzzy-Multi-Objective Decision-Making and feed-forward-Neural Network method. The novelty of this hybrid model is taking into account the weather and meteorology data. The first hybrid model implements Fuzzy-Multi Objective Decision Making as an input data pre-processing technique before the feed-forward Neural Network model. The error statistical indicators of the Fuzzy-Multi Objective Decision Making-Feed Forward Neural Network) model for mean square error is the mean value error 0.001 and root mean square error is mean value error 0.2. Note that the highest root means square error was 4.43 MW and the mean square error was 837.8 MW during the sixteen periods. [ABSTRACT FROM AUTHOR]

This paper proposes a novel methodology for very short term load forecasting of hourly. The proposed methodology is based on meteorology data i.e. temperature, humidity, especially for optimizing the operation of power generating electricity from thermal unit generation. This modelling methodology is a combination of k-nearest neighbor (k-NN) method and feed forward- Neural Network (Feed-Forward-NN) method. The k-NN-Feed-Forward NN model is designed to prediction load for 1 hour ahead based on meteorology data for the target Thermal Unit Generation which position adjacent by twelve hydro thermal unit generation. The novelty of this model is taking into account the meteorology data. A set of load measurement samples was available from the hydro thermal unit generation in Indonesia Region 4 which is used as test data. The first model implements k-NN as a input data preprocessing technique prior to feed forward NN model. The error statistical indicators of k-NN-Feed-Forward-NN method The mean absolute deviation error statistical indicators of k-NN model is 103.48 MW and MAPE is 18.8%. On the other hand, the error statistical indicator for proposed model (Euclidean k-NNfeed forward-NN model) MAD is 19.37 MW and MAPE is 2.21%. Note that the highest mean absolute deviation (MAD) was 75,11 MW and mean absolute percentage error (MAPE) was 10.38% during the twelve period. The models forecasts are then compared to measured data and simulation results indicate that the k-NN-Feed Forward NN-based method presented in this research can calculate hourly load with satisfactory accuracy. [ABSTRACT FROM AUTHOR]