1
SwingNN
User guide
2
Contents
SwingNN ........................................................................................................................... 3
How it Works ................................................................................................................... 4
Getting Started................................................................................................................ 5
Main Windows ................................................................................................................. 6
Grid ................................................................................................................................ 7
Grid Names ............................................................................................................ 10
Network ....................................................................................................................... 12
Importance ................................................................................................................. 14
Graph ........................................................................................................................... 15
Toolbars.......................................................................................................................... 17
Main ............................................................................................................................. 18
View ............................................................................................................................. 21
Tools ............................................................................................................................ 22
Dialogs ............................................................................................................................ 23
Details ......................................................................................................................... 24
Forecast ...................................................................................................................... 26
Progressive Forecast .............................................................................................. 28
Discovered Patterns ................................................................................................ 31
Set Function .............................................................................................................. 33
Set Grid Range.......................................................................................................... 35
Support ....................................................................................................................... 36
Special files ............................................................................................................... 37
Save Grid Details.................................................................................................. 39
3
SwingNN
SwingNN Introduction
Welcome to SwingNN, a powerful and easy to use forecasting application.
Using SwingNN you can:
1.Import text, csv, spreadsheet, image or binary files into the Grid.
2.Use many editing and pre-formatting functions on the grid.
3.Set up any of the grid columns for forecasting.
4.Use the fully automated forecasting facilities.
New Users should do the Getting Started exercises.
Neural Planner Software Ltd
SwingNN
How it Works
Getting Started
Main Windows
Toolbars
Dialogs
4
How it Works
How SwingNN forecasts the future.
The future can never be forecasted with guaranteed accuracy but by using the SwingNN
methods it is possible to achieve good results.
The SwingNN grid is produced by importing txt, csv, xls, bmp or binary files. The grid
can also be produced manually using the editing facilities. Numeric, text and image
columns can be used together in the same grid.
Only future numeric values can be forecasted. Text and image inputs will influence the
forecasting of the numeric values.
A neural network is created from the grid data. Grid input columns are connected to
network input nodes, grid output columns are connected to network output nodes and
grid serial columns are connected to both network input and output nodes. At least one
serial column is needed for forecasting.
The neural network is trained using the grid example rows. The risk level is set to zero.
After the neural network is trained, the serial input values are forced to swing beyond
their limits. The serial output values are forecasted by the neural network. A new neural
network is created and trained using the new inputs and forecasted outputs. The new
neural network is compared with the original neural network. The inputs are adjusted,
another new neural network is created and the risk level is incremented by one. The
process continues until a new neural network agrees with the original one about the
forecasts or the risk level is too high. The forecasts are added to the grid for you to use.
5
Getting Started
The exercises can be started by clicking the Getting Started button on the Tip of the
Day or using the menu command Help > Getting Started...
1. Pressure
In the first exercise you will forecast a future health problem.
The 'Pressure.tvq' sample is a record of a patient's blood pressure taken every day. The
systolic and diastolic pressures vary quite significantly and they do not always rise and
fall together. The patient needs to know if a problem is likely to occur any time in the
near future. The systolic column and diastolic columns are changed to Serial to make the
grid ready to allow rows and future values to be added and tested. 'Action > Forecast'
opens the 'Forecast' dialog. New rows are generated with the forecasted values. The grid
scrolls to show the new rows. The graph of the forecasted systolic values show a future
high blood pressure problem.
2. Random
This exercise shows that SwingNN can find trends and forecast future values even if the
data is random. All the columns are already set to 'Serial' so future values can be
forecasted. When forecasting has finished the risk values are indicated. The forecast
value may be wrong for any risk. That is to be expected when the initial values are
random. SwingNN shows you the risk.
3. Grocer
Every day a retail grocer collects his stock from the wholesale market and sets the prices
he will charge. SwingNN is used to forecast the prices that he will be charging in the next
few days. All the columns are set to 'Input'. The columns for the food prices that are to be
forecast to changed to 'Serial'. The 'Forecast' toolbar icon is enabled and is clicked to
open the 'Forecast' dialog. 'Values to forecast' is changed to 5 and forecasting started.
When forecasting has finished the next days food prices may show some low risk
bargains that could be used as sales leaders without any serious impact on the grocers
profit.
4. Shares
Many applications use neural networks to forecast the price of shares. The forecasts are
often less useful than those of stock market experts. This exercise shows that the forecast
produced by SwingNN is a downward price trend. The expert forecasted a similar
downward trend.
6
Main Windows
Main Windows
Grid
Network
Importance
Graph
7
Grid
The Grid view shows all the Examples arranged in rows and all the Input/Outputs
arranged in columns. The first column contains the Example types and names. The first
row contains the Input/Output types and names. Everything on the Grid can be edited by
moving to the cell containing the value and then pressing the enter key to start the Edit
Grid dialog. The cell can be selected either using the arrow keys or the mouse. A single
click will select the cell and a double click will start the Edit Grid dialog. A double click
on the Example name cell will select the whole row and a double click on the
Input/Output name cell will select the whole column. The row or the column can be
deselected by pressing the Esc key.
Creating a New Grid.
A new Grid is created by pressing the new toolbar button or using the File > New menu
command. The new Grid will be empty except for a horizontal line, a vertical line and an
underline marker that shows the current position in the Grid. New Grid rows and columns
are created at the current position.
Creating the first Example row and Input/Output column
8
Press return and a prompt will appear that says "Create new Example row?". Answer
Yes. Another prompt will appear that says "Create new Input/Output column?". Answer
Yes again. You will now see that the Grid has one cell containing "?", a row name
containing "T:0" and a column name containing "I:0". The "?" indicates that the cell has
no value, the "T:0" indicates that it is a Training Example in row 0 and the "I:0" indicates
that is is an Input in column 0. Press return again and an Edit Grid dialog box will appear
that allows you to enter the cell value. Using the same dialog you can change the
Input/Output column name, mode and type. The dialog can also be used to change the
Example row name and type.
Creating more Input/Output columns.
Move the marker one cell to the right by pressing the right arrow or tab key. Now press
the return key and a prompt will appear that says "Create new Input/Output column?".
Answer Yes. Press return again and the Edit Grid dialog box will appear. This time the
Example row will already contain a name and type. You can set the cell value, the
Input/Output name and the type can be set to "I:" for input, "O:" for output, "X:" for
exclude or "S:" for serial. The mode can be set to "Real", "Integer", "Bool" "Text" or
"Image". Any type of Input/Output column can be inserted into the Grid using the
functions on the Insert menu.
Creating more Example rows.
Move the marker one cell down by pressing the down arrow key. Now press the return
key and a prompt will appear that says "Create new Example row?". Answer Yes. Press
return again and the Edit Grid dialog box will appear. This time the Input/Output column
will already contain a name and type. You can set the cell value, Example name and the
type can be set to T: for training, V: for validating, Q: for querying or X: for exclude.
Any type of Example row be inserted into the Grid using the functions on the Insert
menu.
Copying Example rows and Input/Output columns.
Double click on the name to select the whole row or column. Cut will remove the
selected row or column and place it on the clipboard. Copy will place a copy of the
selected row or column on the clipboard. Paste will insert a copy of the clipboard before
the currently selected row or column. If the clipboard contains a row then Paste will
insert the row into the Grid. If the clipboard contains a column then Paste will insert a
column into the Grid. The invisible Grid data, limits and defaulted values, will be
regenerated after a Paste column thus any neural network that has already been generated
from the Grid will be invalidated.
Rows names
9
The grids rows can have any names. If names are not used the rows are numbered.
Column names
The grid columns can have any names. If names are not used the columns are
numbered.
Grid cells
Grid cells can contain numbers, text or images. The numbers can be integer, real or
boolean.
Grid
Grid Names
10
Grid Names
The meaning of the row and column names.
Serial Column
A serial mode column of Integers. Only serial mode columns are forecasted.
Excluded Columns
Two excluded columns with real numbers and letters. Excluded columns are not
forecasted.
Row Names
11
The original four rows with names.
Swing Forecasts
Four Swing forecasted rows. The rows names show the risk level for each row. If the
name is Forecast it means that there is no risk.
Progressive Forecasts
Progressive forecasts with row names showing the risk and copies of the original row
names. As forecasting continues copies of the forecasts are used for further forecasts.
Cells
The grid cells contain the original and the forecasted values.
12
Network
Neural networks are created when they are required for forecasting. Grid input columns
are connected to network input nodes, grid output columns are connected to network
output nodes and grid serial columns are connected to both network input and output
nodes.
Input node
Input nodes are connected to grid input or serial columns.
Output node
Output nodes are connected to grid output or serial columns.
Hidden node
13
Hidden nodes are connected to input, output or serial nodes.
Connections
Layers of nodes are fully connected to the adjacent layer. A connection can have a
positive or negative weight. The line color shows the polarity and the thickness shows
the weight value relative to the greatest weight.
14
Importance
The importance of each input is the sum of the connection weights from the input node
to the hidden layer.
15
Graph
The column values graph.
Scaled Values
The column values are scaled from 0.0 to 1.0
16
Trend
The trend over the whole graph.
Value && Average
The blue line is the value of every row for the selected column. The red line is the
running average of a tenth of the total rows for the selected column.
Row Number or Name
Rows are numbered or named. The names of the rows can be selected on the right click
menu.
Real Values
The real column values.
17
Toolbars
Toolbars
Main
View
Tools
18
Main
The main toolbar is for file access, grid editing and forecasting.
New
New creates an empty Grid.
Open
Open produces the standard dialog for file navigation and reading.
Import
Import produces the standard dialog for file selection. The file selected produces an
import dialog appropriate to the file type.
Save
Save produces the standard file selection and save dialog.
Print
Print the current view.
About
About displays the application version and registration status.
Cut
19
Cuts the selected row or column to the clipboard.
Copy
Copies the selected row or column to the clipboard.
Paste
Pastes the clipboard contents to the row or column.
Refresh
Refreshes the current view.
Check
Checks that the grid can be used for forecasting.
Swing
Swing Forecast opens the forecasting dialog to set the serial columns to allow input
and output values to swing beyond their limits.
Stop
Stops forecasting.
Pattern
Pattern forecasting
Progressive
Progressive Forecast opens the progressive dialog so that the ranges to use for
forecasting can be set.
Discover
Discover searches for patterns in the grid and then opens the Discovered Patterns
dialog to allow patterns to be used.
Delete latest
Delete Latest deletes the latest copied or forecasted row and cancels the changes to
20
the network used while forecasting. If all forecasted rows are deleted the network is also
deleted.
Delete copies
Deletes all the copied rows.
Delete forecasts
Deletes all the forecasted rows.
21
View
The view toolbar.
Information
Information about the open file.
Grid
Grid shows the grid with scroll bars if needed.
Network
Network shows the latest created neural network.
Graph
Graph shows the select column values graph.
Memory
Memory remembers the currently selected column so it can be compared with any other
column.
Importance
Importance shows the importance of each input.
Zoom
Zoom the current view in or out.
22
Tools
Special Files
Special Files opens the Special Files dialog
Email
Email opens the Support dialog.
23
Dialogs
Dialogs
Details
Forecast
Progressive Forecast
Discovered Patterns
Set Function
Set Grid Range
Support
Special files
24
Details
The Details dialog. All the details of the SwingNN file.
History button
25
Show a list of previous file open and save dates and times.
Save button
Save the SwingNN file details to a text file.
Refresh button
Refresh all the information.
Continue button
Close the information dialog and continue using this SwingNN file.
26
Forecast
The Forecast dialog is used to forecast the values of any serial mode columns.
Forecast type
The two main types of forecast.
Only Serial columns can be used to produce forecasts.
Swing. Includes the number of rows and forecast period.
Pattern. Includes the range of the pattern found.
Rows to forecast
27
Enter the maximum number of rows to forecast.
Risk
Enter the maximum acceptable risk level for the forecasts. SwingNN may go beyond
this level as it tries to reduce the risk.
Rows to include
Check all the row types to include in the forecast.
Limits
Check to allow the forecasted values to go above or below the limits of the current
range.
OK button
Press OK to accept all the settings and close the dialog.
Cancel button
Press Cancel to reject all the settings and close the dialog.
28
Progressive Forecast
An initial network is produced and trained using every row from the First to the Last.
Another network is produced and trained using the rows from the Start and then with all
the following rows up to the maximum. A copy of the row being forecast and a new row
are added to the end of the grid. The new row has the forecasted values. This allows the
forecasted values to be compared with the actual values. More than one forecast can be
produced at every step. The process continues until the maximum steps is reached.
When the process stops the grid is returned to its original state without removing the
added rows. A secondary copy of the grid can be saved along with the secondary
network.
Total Rows
29
The total number of rows in the grid. All rows are counted.
First Training Row
The first row in the grid to use for training the primary network.
Last Training Row
The last row in the grid to use for training the primary network.
Start Forecasting Row
The first row to use for forecasting.
Forecasts per step
The number of forecasts to be produced after each step.
Produce Secondary Network
Check to produce a secondary network.
Secondary Network Name
The file name of the secondary network.
Max rows
The maximum number of rows to use for training at each step.
Max steps
The maximum number of steps.
Risk
30
Forecasting will stop if the risk level is exceeded and cannot be reduced.
Ignore stops
Check to ignore all stops during progressive forecasting.
Defaults button
Click Defaults to set all the dialog controls to their defaults settings for the grid.
OK button
Click OK to accept all the settings and close the dialog.
Cancel button
Click Cancel to reject all the settings and close the dialog.
31
Discovered Patterns
The Discovered Patterns dialog is produced after Action > Discover Patterns has run.
The dialog will show any of the patterns that have been discovered in the grid and can
be used to extend or modify the grid with the discoveries.
Set All buttons
The mode of all the grid columns is set to Serial or Exclude.
Pattern Selection
Press the buttons to find and select a discovered pattern.
Selected Pattern
32
The grid start row, end row and column in the currently selected pattern.
Trend
The trend of the values in the currently selected pattern.
Toggle button
The serial mode of the selected column is inverted.
Include in Forecast button
The selected pattern will be included in any forecasts. Every pattern is included by
default.
Exclude from Forecast button
The selected pattern will be excluded for all forecasts.
Copy to End of Grid button
The selected pattern is copied to the end of the grid.
33
Set Function
The Set Function dialog is opened using the grid, right click, context menu. Functions
can be used on cells, rows, columns or entire grid ranges.
Function
The function to execute.
34
Fail
Stop the function if any failures occur.
OK button
Press OK to accept all the settings and close the dialog.
Cancel
Press the Cancel to reject all the settings and close the dialog.
35
Set Grid Range
The Set Grid Range dialog is opened using the grid, right click, context menu. Functions
can be limited to grid ranges.
Rows
The range of rows for the function.
Columns
The range of columns for the function.
OK button
Press OK to accept the setting and close the dialog.
Cancel button
Press Cancel to reject the settings and close the dialog.
36
Support
FAQ
Press the FAQ button to read the online frequently asked questions.
Emial
Press the email button to send an email to support.
37
Special files
Network and grid components are saved to txt or csv files.
What to save
Check the item that are to be saved.
Extension
Select the file type.
When to save
Select when to save.
OK button
38
Press OK to accept the setting and close the dialog.
Cancel button
Press Cancel to reject the setting and close the dialog.
Special files
Save Grid Details
39
Save Grid Details
The details to save in the Grid special file.
All Example Rows
Check to save all example rows including copies and forecasts.
Forecasted Rows
Check to save the forecast example rows.
Copied Rows
Check to save the copies of example rows.
Row Number
Check to prefix the example row names with the row numbers.
Min/Max Column Details
40
Check to save the minimum and maximum column details at the end of the grid file.
OK button
Click to accept all the settings and close the dialog.
Cancel button
Click to reject all the settings and close the dialog.
41