Using the Data Mining Add-In for Microsoft Office - Microsoft Business Intelligence Tools for Excel Analysts - page 272

Figure 14-8: Results of the Fill From Example report.
The output is a Pattern Report (see Figure 14-8), which shows column and value states as they relate
to the sample values column (High Value Customer). This report orders value states by relative
impact. It's useful to consider the results by each factor; for example, by considering the favors “Yes”
states separately from the favors “No” states. A quick glance shows you that customers located in
Europe with relatively short commute distances (2-5 miles) favors a high value customer status, while
for customers in the Pacific, higher commute distances (5-10 miles) or very short commute distances
(1-2 miles) do not favor high value customer status. Other factors can quickly be identified.
Forecast
The Forecast tool is similar to the Fill From Example tool, but with two important distinctions:
➤ The Forecast tool assumes a time-oriented dataset. The time column could be time of day,
date, or a combination of the two (date + time). Why does it matter that one of the columns
is specifically a date/time value? Because the underlying algorithm, Microsoft Time Series, is
based on a time-ordered dataset.
➤ Where the Fill From Example tool fills in missing values for a single column based on relation-
ships detected with other input columns (column-based pattern), the Forecast tool fills in
missing rows for multiple columns based on relationships detected in other rows (row-based
pattern).