Details on Different Graph Types

Top  Previous  Next

_bm50Variable Graphs Detail

This module allows you to graph one or more of the network's variables in many different ways.  When putting data into the network, you might want to create a scatter plot of two variables to insure that the training patterns are representative of the entire problem domain.  Once the data has been processed by the network, you can graph the network's predictions compared to the actual answers.



1. Graph Variable(s) Across all Patterns. Graph one or more variables, even if the variables are different types, across all patterns in the file.  This graph is useful for analyzing time series data.  You also have the option of graphing variables across a selected number of patterns.



2. Graph Variable Sets in a Pattern. Use this graph if all of the variables in a pattern are of the same type, such as 100 points in a physiological signal like an electrocardiogram.   The graph will display what the electrocardiogram looks like.



3. Correlation Scatter Plot. Graph one variable against another for all patterns, providing a correlation factor.  This graph displays the linear correlation coefficient between the two graphed variables.



4. High Low Close Graph. Graph the high, low, and close variables used for market predictions.