IMPLEMENTASI METODE ENSEMBLE K-NEAREST NEIGHBOR UNTUK PREDIKSI JUMLAH UANG BEREDAR
Daftar Isi:
- Predicting money supply is needed in order to anticipate future inflation. Predicting is done with a modified kNN regression and a combination of the final steps using the ensemble technique, namely weighted means. The data used uses data on the money supply and variables that affect the money supply, namely foreign assets, central government bills, and other securities bills. These variables are used to predict the money supply that will be predicted. The data used monthly money supply data in 2014 - 2018 to predict the money supply in 2016, 2017,2018 with training data 2 years before. In this study, the best evaluation value obtained is the accuracy in predicting predictions in 2016 k-9 with MAE = 47.59946, MAPE = 1.0056341, and RMSEP = 59.847633 for ensemble with MAE = 48.394093, MAPE = 1.0245531 and RMSEP = 59.975723, 2017 predictions k-3 with MAE = 125,575, MAPE = 2.390784, and RMSEP = 148.4847 for ensemble with 171.6709, MAPE = 3.283542 and RMSEP = 189.6373, and predictions for 2018 k-10 with MAE = 37.34648, MAPE = 0.672266, and RMSEP = 42.485 for 4284 ensemble with MAE = 40.38839, MAPE = 0.728408 and RMSEP = 45.01614.