| 摘要 |
A method, system and apparatus for identifying malfunctioning remote devices, such as point of sale devices, are disclosed. In operation, a performance history for one or more remote devices is built and a normal operating day is determined from the performance history. Also extracted from this performance history are average and variance values for various aspects of satisfactory remote device performance. During remote device evaluations, the current operating performance of each remote device is compared with the normal day and the average and variance values for the respective remote device. For remote devices operating outside of the normal day, average and variances values, a likelihood estimate that the remote device is actually malfunctioning is calculated. For remote devices determined likely to be malfunctioning, an approximation of the remote device's losses are made such that the remote device with the greatest losses may be prioritized for maintenance or other purposes.
|
