org.encog.app.analyst
Class EncogAnalyst

java.lang.Object
  extended by org.encog.app.analyst.EncogAnalyst

public class EncogAnalyst
extends Object

The Encog Analyst runs Encog Analyst Script files (EGA) to perform many common machine learning tasks. It is very much like Maven or ANT for Encog. Encog analyst files are made up of configuration information and tasks. Tasks are series of commands that make use of the configuration information to process CSV files.


Field Summary
static String TASK_FULL
          The name of the task that SHOULD everything.
static int UPDATE_TIME
          The update time for a download.
 
Constructor Summary
EncogAnalyst()
          Construct the Encog analyst.
 
Method Summary
 void addAnalystListener(AnalystListener listener)
          Add a listener.
 void addCommand(Cmd cmd)
          Add a command.
 void analyze(File file, boolean headers, AnalystFileFormat format)
          Analyze the specified file.
 int determineInputCount()
          Determine the input count.
 int determineInputFieldCount()
          Determine the input field count, the fields are higher-level than columns.
 int determineMaxTimeSlice()
           
 int determineMinTimeSlice()
           
 int determineOutputCount()
          Determine the output count, this is the number of output columns needed.
 int determineOutputFieldCount()
          Determine the number of output fields.
 int determineTotalColumns()
           
 int determineTotalInputFieldCount()
          Determine the total input field count, minus ignored fields.
 int determineUniqueColumns()
          Determine how many unique columns there are.
 int determineUniqueInputFieldCount()
          Determine the unique input field count.
 int determineUniqueOutputFieldCount()
          Determine the unique output field count.
 void download()
          Download a raw file from the Internet.
 void executeTask(AnalystTask task)
          Execute a task.
 void executeTask(String name)
          Execute a task.
 int getLagDepth()
           
 int getLeadDepth()
           
 List<AnalystListener> getListeners()
           
 int getMaxIteration()
           
 MLMethod getMethod()
           
 Map<String,String> getRevertData()
           
 AnalystScript getScript()
           
 AnalystUtility getUtility()
           
 boolean isTimeSeries()
           
 void load(File file)
          Load the specified script file.
 void load(InputStream stream)
          Load from an input stream.
 void load(String filename)
          Load from the specified filename.
 void reanalyze(File file, boolean headers, AnalystFileFormat format)
          Analyze the specified file.
 void removeAnalystListener(AnalystListener listener)
          Remove a listener.
 void reportTraining(MLTrain train)
          Report training.
 void reportTrainingBegin()
          Report that training has begun.
 void reportTrainingEnd()
          Report that training has ended.
 void save(File file)
          Save the script to a file.
 void save(OutputStream stream)
          Save the script to a stream.
 void save(String filename)
          Save the script to a filename.
 void setCurrentQuantTask(QuantTask task)
          Set the current task.
 void setMaxIteration(int i)
          Set the max iterations.
 void setMethod(MLMethod method)
           
 boolean shouldStopCommand()
          Should the current command be stopped.
 void stopCurrentTask()
          Stop the current task.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

TASK_FULL

public static final String TASK_FULL
The name of the task that SHOULD everything.

See Also:
Constant Field Values

UPDATE_TIME

public static final int UPDATE_TIME
The update time for a download.

See Also:
Constant Field Values
Constructor Detail

EncogAnalyst

public EncogAnalyst()
Construct the Encog analyst.

Method Detail

addAnalystListener

public void addAnalystListener(AnalystListener listener)
Add a listener.

Parameters:
listener - The listener to add.

addCommand

public void addCommand(Cmd cmd)
Add a command.

Parameters:
cmd - The command to add.

analyze

public void analyze(File file,
                    boolean headers,
                    AnalystFileFormat format)
Analyze the specified file. Used by the wizard.

Parameters:
file - The file to analyze.
headers - True if headers are present.
format - The format of the file.

reanalyze

public void reanalyze(File file,
                      boolean headers,
                      AnalystFileFormat format)
Analyze the specified file. Used by the wizard.

Parameters:
file - The file to analyze.
headers - True if headers are present.
format - The format of the file.

determineInputCount

public int determineInputCount()
Determine the input count. This is the actual number of columns.

Returns:
The input count.

determineInputFieldCount

public int determineInputFieldCount()
Determine the input field count, the fields are higher-level than columns.

Returns:
The input field count.

determineOutputCount

public int determineOutputCount()
Determine the output count, this is the number of output columns needed.

Returns:
The output count.

determineOutputFieldCount

public int determineOutputFieldCount()
Determine the number of output fields. Fields are higher level than columns.

Returns:
The output field count.

determineUniqueColumns

public int determineUniqueColumns()
Determine how many unique columns there are. Timeslices are not counted multiple times.

Returns:
The number of columns.

determineUniqueInputFieldCount

public int determineUniqueInputFieldCount()
Determine the unique input field count. Timeslices are not counted multiple times.

Returns:
The number of unique input fields.

determineTotalInputFieldCount

public int determineTotalInputFieldCount()
Determine the total input field count, minus ignored fields.

Returns:
The number of unique input fields.

determineUniqueOutputFieldCount

public int determineUniqueOutputFieldCount()
Determine the unique output field count. Do not count timeslices multiple times.

Returns:
The unique output field count.

download

public void download()
Download a raw file from the Internet.


executeTask

public void executeTask(AnalystTask task)
Execute a task.

Parameters:
task - The task to execute.

executeTask

public void executeTask(String name)
Execute a task.

Parameters:
name - The name of the task to execute.

getLagDepth

public int getLagDepth()
Returns:
The lag depth.

getLeadDepth

public int getLeadDepth()
Returns:
The lead depth.

getListeners

public List<AnalystListener> getListeners()
Returns:
the listeners

getMaxIteration

public int getMaxIteration()
Returns:
The max iterations.

getRevertData

public Map<String,String> getRevertData()
Returns:
The reverted data.

getScript

public AnalystScript getScript()
Returns:
the script

load

public void load(File file)
Load the specified script file.

Parameters:
file - The file to load.

load

public void load(InputStream stream)
Load from an input stream.

Parameters:
stream - The stream to load from.

load

public void load(String filename)
Load from the specified filename.

Parameters:
filename - The filename to load from.

removeAnalystListener

public void removeAnalystListener(AnalystListener listener)
Remove a listener.

Parameters:
listener - The listener to remove.

reportTraining

public void reportTraining(MLTrain train)
Report training.

Parameters:
train - The trainer.

reportTrainingBegin

public void reportTrainingBegin()
Report that training has begun.


reportTrainingEnd

public void reportTrainingEnd()
Report that training has ended.


save

public void save(File file)
Save the script to a file.

Parameters:
file - The file to save to.

save

public void save(OutputStream stream)
Save the script to a stream.

Parameters:
stream - The stream to save to.

save

public void save(String filename)
Save the script to a filename.

Parameters:
filename - The filename to save to.

setCurrentQuantTask

public void setCurrentQuantTask(QuantTask task)
Set the current task.

Parameters:
task - The current task.

setMaxIteration

public void setMaxIteration(int i)
Set the max iterations.

Parameters:
i - The value for max iterations.

shouldStopCommand

public boolean shouldStopCommand()
Should the current command be stopped.

Returns:
True if the current command should be stopped.

stopCurrentTask

public void stopCurrentTask()
Stop the current task.


isTimeSeries

public boolean isTimeSeries()
Returns:
True, if any field has a time slice.

getMethod

public MLMethod getMethod()
Returns:
the method

setMethod

public void setMethod(MLMethod method)
Parameters:
method - the method to set

determineTotalColumns

public int determineTotalColumns()

determineMaxTimeSlice

public int determineMaxTimeSlice()

determineMinTimeSlice

public int determineMinTimeSlice()

getUtility

public AnalystUtility getUtility()
Returns:
the utility


Copyright © 2014. All Rights Reserved.