In this report, Leading Economic Indicator (LEI) data
and Gross Domestic Product (GDP) data have been
analyzed to determine if changes in the ten indicators
can be used to predict changes in GDP. Three neural
network methods and one statistical method were used
to complete the analysis. For this project, the intent
was to use multiple regression and backpropagation to
develop correlations in which LEI values are used to
predict the GDP change in the following quarter.
Alternatively, Kohonen's self-organizing map and
hierarchical clustering were used to group months of
LEI data into clusters to determine if months in a
cluster (and thus months with similar LEI values) also
have similar changes in GDP.