NN5_Complete_107 data set 1: Description. These series are part of the so-called reduced dataset of the Neural Network Forecasting Competition 5 (NN5). The data are daily cash withdrawal amounts at different ATMs in the UK, measured over a period of 2 years. The reduced dataset contains 11 series in total. The missing values that are present in the original version of the series were removed by a method proposed for this dataset in: J.D. Wichard. Forecasting the NN5 time series with hybrid models. International Journal of Forecasting, 27(3):700–707, 2011. http://www.neural-forecasting-competition.com/NN5/ 2: Type. Time series 3: Origin. Real world 4: Instances. 782 5: Features. 4 6: Missing values. No 7: Header. @relation NN5_Complete_107 @attribute TimeStamp real [163, 790] @attribute V1/-4 real [1.84240362811791, 41.5958049886621] @attribute V1/-3 real [1.84240362811791, 41.5958049886621] @attribute V1/-2 real [1.84240362811791, 41.5958049886621] @attribute V1/-1 real [1.84240362811791, 41.5958049886621] @attribute V1/0 real [1.84240362811791, 41.5958049886621] @inputs V1/-4, V1/-3, V1/-2, V1/-1 @outputs V1/0