Forecasting > Forecast Profiles Overview

Forecast Profiles Overview

Forecast Profiles are sets of configuration options that can be used when generating forecasts.

You can create multiple Forecast Profiles, each storing a different configuration, and use them to generate forecasts for any Environment. Forecast Profiles offer a convenient way to set up a particular configuration once so that it can be used for multiple forecast runs. Each Forecast Profile consists of two primary groups of settings, which are discussed in the following sections.

Creating Forecast Profiles requires specific permissions, and you can always edit profiles you created and view the details of those profiles. Separate permissions are required to:

Edit profiles created by others
View the details of a profile used in a forecast run if you did not create that profile
See profiles created by others in the Forecast Profiles list

Forecast Model Parameters

inContact WFMv2 supports forecasting using a number of different models and expert systems. The settings in this category enable you to select a model and configure the appropriate options, including:

Forecast — you can choose to forecast contact volumes only, average service times only, or both
Forecast Model — see Forecast Models
Box-Cox Power Transform
Forecast Method — see Forecast Methods
Data Analysis Bucket Size — you can choose how data should be aggregated for analysis
Forecast Bucket Size — you can choose how the forecast should be generated and stored
Seasonality — see Seasonality Options

Forecast Models

inContact WFMv2 supports the following models:

Best Pick — Evaluates the Exponential Smoothing and Box-Jenkins ARIMA models to determine the structure and parameters for these models, compare their accuracies, and select the one that has the highest accuracy to generate forecasts. Best Pick is the default Forecast Model.
Exponential Smoothing — Uses the average of historical data with exponentially decreasing multipliers. In its seasonal version, this model uses three related equations for level, trend, and seasonality as the basis for a forecast. inContact WFMv2 provides simple Exponential Smoothing, Double Exponential Smoothing, Triple Exponential Smoothing - Multiplicative, and Triple Exponential Smoothing - Additive.
Box-Jenkins ARIMA — Time series components used in Box-Jenkins methodology are auto-regressive (AR) and moving averages (MA) of errors. They are commonly called ARIMA (Auto-Regressive Integrated Moving Average) models. ARIMA models use either past values (the auto-regressive model), past errors (the moving average model) or combinations of past values and past errors (the ARIMA model). The Box-Jenkins method is well suited to handle complex time series and other forecasting situations in which the basic pattern is not readily apparent. It uses an iterative approach to identify a useful model from a general class of models. The chosen model is then checked against the actual data to see if it accurately describes the series. If the model does not fit well, the process is repeated until the most accurate model is found.
Curve Fitting — Also known as Multilinear Seasonal Regression. Generally used to model the trend and seasonality in a data set, this model uses dummy variables in a multiple linear regression model for seasonal effects. Thus, Multilinear Seasonal Regression treats seasonal effects as additive rather than multiplicative.

Forecast Methods

inContact WFMv2 supports forecasting by:

Day of Week — Contact volumes and service times for the past Mondays are used for forecasting contact volumes and service times for future Mondays, contact volumes and service times for the past Tuesdays are used for forecasting contact volumes and service times for future Tuesdays, and so on. By Day of Week is the default Forecast Method.
Whole Time Series — The entire time series is analyzed together and used to forecast over the forecast horizon.

Seasonality Options

Seasonality represents a periodic pattern of behavior in a time series, like the repeating pattern over the days of a week. inContact WFMv2 takes seasonality explicitly into account during the model fitting process. You must select a bucket size smaller than the seasonality pattern value. Seasonality options include:

No Seasonality — No consideration is given to any seasonal pattern that may exist in the data.
Daily — Considers the daily periodic pattern over a week, month, quarter, or year
Weekly — Considers the weekly periodic pattern over a month, quarter, or year
Monthly — Considers the monthly periodic pattern over a quarter or year
Quarterly — Considers the quarterly periodic pattern over a year

Data Analysis Parameters

The settings in this category enable you to control how inContact WFMv2 analyzes the data when creating forecasts. These options allow you to: