Loading...
 

Neural network

Neural network

Description

A business object of the neural network type refers to a model previously created with the Build Neural Network app for mapping input data to output data. The number and type of input and output data was previously trained in the model and can be used directly in the maintenance app.

Functionality

Input window

This app is used to maintain and query a neural network. The core of the neural network is the reference to the actual neural network saved as an .xml file (see Storage field). The other data fields are used to further describe and classify the neural network.

This app can also be used to test the neural network. To do this, you must enter the input data defined by the model; after pressing the "Apply" button, the most probable output values are displayed.

Menu
Menu item Description
Edit -
: search terms Maintain the search terms under which a neural network is to be found.
: notes Create/edit notes.
: Create transaction note -
: : note Create transaction document.
: operations Call up processes.
: validity Set validity.
: Close The app is closed.
Search Call up list window
Preferences Call up options window
Toolbar
Icon Description
new.png Create new neural network
save.png Save changes
delete.png Delete neural network
pin.pngpinned.png Create/edit notes
CX_PROCEEDINGS.png Call up operations
CX_PROCEEDINGS_new.png Create operation document
CX_VALIDITY.png Set validity
Fields
Field Description
Neural network
: data -
: : number Unique number of the neural network. This number is optional.
: : short name* Short name of the neural network. This field must be maintained.
: : classification Optional assignment to the classification term "NEURAL_NETWORK_DECORATOR_CLASSIFICATION".
: : description Free text to describe the neural network.
: : filing File path to the actual model of the neural network saved as an .xml file.
: : search terms Maintenance of the search terms under which a neural network is to be found.
: input data List of input fields for the neural network. If these are filled in and the "Apply" button is pressed, the model determines the most probable values for the output data, which are displayed in the group to the right.
: Output data List of output values determined by the model. The hit rate for the value is displayed directly to the right; the button next to it can be used to display the list of possible other values sorted by hit rate for each value.
Features
: list List of characteristics for further description of the neural network.
: : Create Create new feature.
: : Insert Insert existing feature.
: : Remove Remove feature from the list.
Buttons
Button Description
Action -
: Back The app is closed.
: Delete The displayed neural network is deleted.
: New The input window is cleared.
: Save Changes to the neural network are saved.

 

List window

This window is used to list the neural networks stored in the system.

Menu
Menu item Description
Edit -
: Reset Reset list:
The list and search fields are cleared.
Options Call up options window
Toolbar
Icon Description
CX_AMOUNT_calculation.png Perform multiple search: All fields marked in blue are included in this search. With the multiple search it is possible to search for several criteria at the same time.
list_clear.png Reset list:
The list and search fields are cleared.
new.png Create new neural network
open.png Edit neural network
delete.png Delete neural network
pin.pngpinned.png Create/edit notes
CX_PROCEEDINGS.png Display operations
CX_VALIDITY.png Set validity
copytoclipboard.png Copy all objects to the clipboard
Search fields
Field Description
Number Unique number of the neural network.
Search term Text search for a neural network.
Classification Search for neural network based on the assignment to a classification.
Buttons
Button Description
Filter -
: classification Filtering by order term
: Advanced search Call up advanced search
Action -
: Back The app is closed.
: Delete The selected neural network is deleted.
: New An empty input window is opened.
: Edit The current neural network is displayed in the input window.

 

Selection window

This window is used to select neural networks.

Search fields
Field Description
Number Unique number of the neural network.
Search term Text search for a neural network.
Classification Search for neural network based on the assignment to a classification.
Buttons
Button Description
Filter -
: classification Filtering by order term
: Advanced search Call up advanced search
Action -
: Back The app is closed.
: Select Confirm selection

 

Options window

All settings for the search can be made here.

Menu
Menu item Description
Edit -
: Set default settings all changed settings are reset to default
Fields
Field Description
List options
: Search options -
: : Immediate takeover in case of a hit If only one neural network is found during a search, this is opened in the input window.
: : Clear list before new search The list is cleared before each new search
: : General search Normal search
Search terms from ... use yourself
: : Advanced search So-called wildcards can be used for the search, such as the * or AND OR...
Search terms of the other roles of the ... are taken into account
: validity Validity
: : With exchange object If an invalid object has a valid exchange object, only the valid exchange object is displayed (default).
: : Without exchange object Possibly existing exchange objects are ignored, only the original object is listed. If this original object is invalid, it is only listed if the listing of invalid objects (selection: All objects) has been selected in the further option.
: : Only valid objects Only valid objects are listed (default).
: : All objects Valid and invalid objects are listed.
: Number of objects Determines the maximum number of objects displayed in the list during a search.
: : All objects All objects found are listed
: : Number of objects The number of listed objects can be specified manually.
: : 100 objects Only a maximum of 100 objects are output in the list
Default settings
: Global presets Global presets can be saved for the client and the cost centre. These are displayed here as well as the current default options that each module has. To create a global preset, a user preset must first be saved. Right-clicking on this user setting and selecting the context menu item "Edit user-defined setting" opens an editing window for this setting, on which the assignment from employee / user to a client or cost centre can be changed. ("Level" button)
: User Preferences Each user can save as many presettings as desired. The only thing to note is that if several settings are found, the first one found is used as the default setting. The name of the setting can be changed with a double click.
Buttons
Button Description
OK The settings are saved and the window closes.

 

Related topics

 

Technical documentation

Base module (provider)
Module name

cxNeuralNetwork.mod

Classes

CX_NEURAL_NETWORK
CX_NEURAL_NETWORK_DECORATOR

Editing module

Module name

neuralNetworkEdit.app

Classes

CX_NEURAL_NETWORK
CX_NEURAL_NETWORK_DECORATOR

Security

In addition to restricting access rights via the class and its data fields, the module can be restricted in its use via some of the received messages.

Received messages
Message Parameter Function Security
EDIT_NEURAL_NETWORK      
Sent messages
Message Parameters Function Receive module
NEURAL_NETWORK_CHANGED      
NEURAL_NETWORK_DELETED      

Selection module

Module name

neuralNetworkSelect.app

Classes

CX_NEURAL_NETWORK
CX_NEURAL_NETWORK_DECORATOR

Security

In addition to restricting access rights via the class and its data fields, the module can be restricted in its use via some of the received messages.

Received messages
Message Parameter Function Security
SELECT_NEURAL_NETWORK      
Sent messages
Message Parameters Function Receive module
NEURAL_NETWORK_SELECTED      

Operational business