Members
collectErrorDetails
Collects together a set of name/value pairs representing details of an error, in the case that a call to $http.get() results in an error.
convertDoubleToGreyShadeColour
Converts the inputted numeric value into an html hex color, scaled between upper and lower values defined in weightBlackShadeValue and weightWhiteShadeValue
- Source:
convertErrorValuesToHtml
Converts a set of key/value pairs representing details of an error to an HTML-formatted error message.
- Source:
getCost
Gets the current cost based on the training performed so far from the REST API.
getFontColourForWeightValue
Gets the appropriate font colour corresponding to the inputted weight value
- Source:
getHiddenLayerActivationValues
Gets the activation values of the hidden layer of the network from the REST API.
getHiddenToOutputLayerWeights
Gets the hidden to output layer weights from the REST API.
getInputToHiddenLayerWeights
Gets the input to hidden layer weights from the REST API.
getOutputLayerActivationValue
Gets the activation value of the output layer of the network from the REST API.
getPrediction
Calls the REST API to predict the result data for two input values based on the neural network's current weights.
hideWaitDialog
Closes the wait dialog
- Source:
initializeWaitDialog
Initializes values in the $scope.dataInterfacePostExecute object and shows a wait dialog in preparation for a long-running background process
- Source:
mainController
The main controller for the backpropagationVisualization application.
- Source:
neuralNetworkDataInterfaceService
Provides a data interface to the REST API running the neural network.
populateAllHiddenLayerValues
Retrieves all the weights and hidden layer activation values from the neuralNetworkDataInterfaceService and combines them in to an array.
Allows all of these values to be shown iteratively in a single row of an HTML table using the 'data-ng-repeat' attribute.
- Source:
populateCost
Retrieves the cost value from the neuralNetworkDataInterfaceService and sets it on the scope
- Source:
populateOutputLayerActivationValue
Retrieves the output layer activation value from the neuralNetworkDataInterfaceService and sets it on the scope
- Source:
predict
Retrieves the predicted output value based on the scope prediction input values. Also sets the output value on the scope.
- Source:
resetHiddenLayerActivationValues
Sets the hidden layer activation values on the scope to 0
- Source:
resetWeights
Resets the weights in the neuralNetworkDataInterfaceService and sets them on the scope
- Source:
resetWeights
Resets the weights in the network via the REST API.
setHiddenLayerActivationValues
Sets the hidden layer activation values on the scope
- Source:
setHiddenToOutputLayerWeights
Sets the hidden to output layer weights on the scope
- Source:
setInputToHiddenLayerWeights
Sets the input to hidden layer weights on the scope
- Source:
setLogicalOperator
Sets the logical operator represented by the training data on the REST API.
setLogicalOperator
Sets the training data corresponding to the logical operator defined in $scope.logicalOperator on the neuralNetworkDataInterfaceService
- Source:
showErrorDialog
Displays a modal error dialog
- Source:
showWaitDialog
Displays a modal wait dialog
- Source:
train
Trains the neural network based on training parameters set on the scope. Updates the weight values, activation values, and cost on the scope when completed.
- Source:
train
Calls the REST API to train the neural network.
trimDouble
Trims decimal values to a fixed number of characters
- Source:
utilitiesService
Javascript utilities for the application.
- Source:
Type Definitions
costCallback(cost)
Parameters:
| Name | Type | Description |
|---|---|---|
cost |
number | The cost of training |
errorCallback(data)
Parameters:
| Name | Type | Description |
|---|---|---|
data |
Object | An array of name/value pairs containing the titles and values of items describing an error situation (usually resulting from a failed called to $http.get()) |
hiddenLayerActivationValuesCallback(hiddenLayerActivationValues)
Parameters:
| Name | Type | Description |
|---|---|---|
hiddenLayerActivationValues |
Array.<number> | An array containing the hidden layer activation values |
hiddenToOutputLayerWeightsCallback(hiddenToOutputLayerWeights)
Parameters:
| Name | Type | Description |
|---|---|---|
hiddenToOutputLayerWeights |
Array.<number> | An array containing the hidden to output layer weights |
inputToHiddenLayerWeightsCallback(inputToHiddenLayerWeights)
Parameters:
| Name | Type | Description |
|---|---|---|
inputToHiddenLayerWeights |
Array.<Array.<number>> | An array of arrays of numbers containing the input to hidden layer weights |
outputLayerActivationValueCallback(outputLayerActivationValue)
Parameters:
| Name | Type | Description |
|---|---|---|
outputLayerActivationValue |
number | The output layer activation value |
predictionCallback(predictionResult)
Parameters:
| Name | Type | Description |
|---|---|---|
predictionResult |
number | The value of a prediction (between 0 and 1) |
successCallback(data)
Parameters:
| Name | Type | Description |
|---|---|---|
data |
Object | The data object from the response object passed back from a successful call to $http.get() |