Skip to main content

Advertisement

Table 2 Example of a PreCount feature matrix divided into input and target feature sets

From: PreCount: a predictive model for correcting real-time occupancy count data

Days Input feature set Target feature set
  Calendar Period of day Raw count Count error
d j d n j d t j w e j h j \(pd_{d_{j}, s_{0}}\) \(pd_{d_{j}, s_{1}}\) \(pd_{d_{j}, s_{k}}\) \(CC_{r_{d_{j},s_{0}}}\) \(CC_{r_{d_{j},s_{1}}}\) \(CC_{r_{d_{j},s_{k}}}\) \(CC_{e_{d_{j},s_{k-10}}}\) \(CC_{e_{d_{j},s_{k-9}}}\) \(CC_{e_{d_{j},s_{k}}}\)
2016-08-28 0 0 3 0 0 0 3 0 57 12 0 -12 11
2016-09-01 1 1 3 0 0 4 3 -2 65 32 -7 0 2
2016-10-18 2 1 0 1 0 4 4 5 88 25 4 2 23
Today 3 1 0 1 0 4 3 -3 53 27 ? ? ?