Forecasting is the first important stage of workforce management planning. WFM forecastors create the forecast of ,including others, volume of conversations to be expected in some future time. Forecasting future contacts is a difficult task because of the uncertainty associated with the many factors that determine contacts as well as the nature of distribution of contacts in time.
Time series data is an object observed in many consecutive units of time in chronological order.
While there is not a single best method for forecasting, there are widely used methods and best practices. We will discuss some of the most commonly used time series models and hybrid approaches in this blog. We will also cover the metrics which can indicate the accuracy of the forecast.
accuracy() function from forecast package gives mean error (ME), root mean square error (RMSE), mean absolute error (MAE), mean percentage error (MPE), mean absolute percentage (MAPE), mean absolute squared error (MASE), autocorrelation function index (ACFI) and Theil’s U values.
MAPE is considered the best accuracy measure since it is not sensitive to sign of error and the magnitude of units. We will use MAPE measures to compare the accuracy of different methods to be discussed.
The html output is currently available at rPubs: https://rpubs.com/Tesfahun_Boshe/873021