Subsections

Overview

SciCraft have 76 different nodes which are divided into 6 categories; file, input , plot, toolboxes, tools and view. Each of the function nodes in toolboxes, which are linked up with the plug-ins in contrib and core makes use of an external program when they run.

File

These nodes (table 3.2 and 3.3) are capable of reading and writing plug-in specific files.

File type support

Table 3.1: File type support
Input node Output node
Matlab R
BASE Molecule(PDB)
Octave Octave
R
Molecule(PDB)
Textfile
GaussDal
Experiment



Table 3.2: Summary of the InputNode.
Name InputNode
Category FileHandler
Description Node that reads data from a file.
Usage This node is used for retrieving data from a file into a Module Diagram. A file is read and the content in this file is parsed into data which can be used by SciCraft.
Parameters Filename
Input ports None
Output ports One port for each variable in the current file will be added.



Table 3.3: Summary of the OutputNode.
Name OutputNode
Category Filehandlers
Description Node that writes data to a file.
Usage This node is used for writing results calculated by SciCraft to a file. Only file formats supported by SciCraft can be written.
Parameters Filename
Input ports One port for each variable that will be written to the file.
Output ports None


Input

These nodes (table 3.4 and 3.5) are capable of reading input commands for SciCraft.


Table 3.4: Summary of the ParamNode.
Name Param
Category Input
Description Node that reads commands from the user.
Usage This node is used for sending commands to SciCraft.
Parameters Command
Input ports None
Output ports Data



Table 3.5: Summary of the UserQuery.
Name UserQuery
Category Input
Description Node that reads queries from the user.
Usage Use this node to query the system.
Parameters Command
Input ports None.
Output ports Data


Plot nodes

Nodes capable of showing plots based on data in SciCraft. These nodes are summed up in (table 3.6 and 3.7) and a detailed coverage of plotting is given in section [*].


Table 3.6: Summary of the Plot nodes.
Name PlotNode
Category Plot
Description Node that is used to make plot based on calculated data in the Module Diagram.
Usage This node's task is to show plots bases on data coming in on the plotnode. In later edition of SciCraft it will be possible to select elements in the plot and then send the data and a command to an edit node to remove these elements from the dataset.
Parameters None
Input ports Arbitrarily many ports with input matrices
Output ports In later editions there will be one exit ports where a command string is set. These ports will then be connected with an edit node which will process the current dataset.



Table 3.7: Summary of the DevelPlot nodes.
Name DevelPlot
Category Plot
Description PlotNode under development. Better plots, but less functionallity.
Usage This node's task is to show plots bases on data coming in on the userplot node.
Parameters None
Input ports Arbitrarily many ports with input matrices
Output ports In later editions there will be one exit ports where a command string is set. These ports will then be connected with an edit node which will process the current dataset.


Toolboxes/function nodes

Plugin nodes which executes different functions in an external program, e.g. Octave or R.

Table 3.8: Summary of the Function nodes.
Name FunctionNode
Category None
Description Node responsible for communication with external programs.
Usage This node is used for communication with external programs. SciCraft contains scripts which are used to generate different function nodes. These nodes will then use these scripts to process data through the external program.
Parameters Depending on the type of script the node contains. None of these are meant to be edited by the user.
Input ports One port for each variable needed by the script this node will execute.
Output ports One port for each data element generated by the external program.


Contrib

Classification

Kmeansc node

The node performs k-means cluster analysis. Table 3.9 gives a summary of the Kmeansc node.


Table 3.9: Summary of the Kmeansc node.
Name Kmeansc
Category Contrib/Classification
Description Node performs k-means cluster analysis.
Usage Add input matrix, number of clusters to search for, and maximum number of iterations.
Parameters Command
Input ports Input matrix, numbers of clusters and max iterations.
Output ports An exit port with CNT/R object with cluster centers.


Knn2 node

The node performs k-nearest neighbour classification. Table 3.10 gives a summary of the Knn2 node.


Table 3.10: Summary of the Knn2 node.
Name Knn2
Category Contrib/Classification
Description Node performs k-Nearest neighbour classification.
Usage Add data matrix as the independent variables for the calibration samples. Add a vector containing the classes for the samples given in the matrix. A value for the number of neighbours considered. An optinal matrix with a set of test cases can be added. .
Parameters Command
Input ports Input matrix, A vector containing classes, number of neighbours, test matrix.
Output ports Percentwise calibration error, prediction error and predicted classes.


lda2 node

The node performs linear discriminant analysis. Table 3.11 gives a summary of the lda2 node.


Table 3.11: Summary of the lda2 node.
Name lda2
Category Contrib/Classification
Description Node performs linear discriminant analysis.
Usage Add a matrix with values for the calibration and a vector containing the classes for the samples given in the matrix.
Parameters Command
Input ports Input matrix, vector containing the different classes.
Output ports R object with the estimated classification rule. The percentwise calibration error and prediction error.


Pcalda node

The node performs a linear discriminant analysis using the scores from a principal component analysis. The optimal number of components is found using crossvalidation. Table 3.12 gives a summary of the pcalda node.


Table 3.12: Summary of the pcalda node.
Name Pcalda
Category Contrib/Classification
Description Node performs lda on pca data.
Usage Add a matrix containing values for the calibration and a vector containing different classes.
Parameters Command
Input ports Input matrix, input vector.
Output ports R object with estimated classification rule, percentwise calibration error and prediction error, loadings from the pca.


Pcapred node

The node performs prediction of classes using principal analysis as preprocessing to linear or quadratic discrimination analysis. Table 3.13 gives a summary of the pcapred node.


Table 3.13: Summary of the Pcapred node.
Name Pcapred
Category Contrib/Classification
Description Node performs prediction of classes based on pca before linear or quadratic discrimination analysis.
Usage An Pcapred node makes use of a rule from either the pcalda or pcaqda node, the loadings from either pcalda orpcaqda and matrix with the independent variables.
Parameters Command
Input ports Classification rule, pcalda or pcaqda loadings, matrix.
Output ports Predicted classes.


Pcaqda node

The node performs quadratic discriminant analysis using the scores from a principal component analysis. The optimal number of components is found using crossvalidation.. Table 3.14 gives a summary of the pcaqda node.


Table 3.14: Summary of the Pcaqda node.
Name pcaqda
Category Contrib/Classification
Description Node performs quadratic discriminant analysis using scores from a pca.
Usage The pcaqda operates on a matrix with indepentdent variables for the calibration, with a vector containing the classes for the samples given in the matrix.
Parameters Command
Input ports Input matrix, and vector containign the classes.
Output ports R object, precentwise clibration error, prediction error and loadings from the pca.


Pred2 node

The node performs prediction of classes using linear or quadratic discrimination analysis. Table 3.15 gives a summary of the pred2 node.


Table 3.15: Summary of the Pred2 node.
Name Pred2
Category Contrib/Classification
Description Node performs prediction of classes using linear or quadratic discrimination analysis.
Usage Add classification rule from either lda2 or qda2 node, and a matrix with independent variables as the sample with unknown classes.
Parameters Command
Input ports lda2 or qda2 rule, input matrix.
Output ports The predicted classes for the unknown samples.


Qda2 node

The node performs quadratic discriminant analysis. Table 3.16 gives a summary of the qda2 node.


Table 3.16: Summary of the qda2 node.
Name qda2
Category Contrib/Classification
Description Node performs quadratic discriminant analysis.
Usage Add a matrix with independent variables for classification, a vector containing classs for the sample and a test set.
Parameters Command
Input ports Input matrix, vector with classes, test set.
Output ports R object with est. class. rule, percentwise calibration error, prediction error and list of predicted classes from the test data.


Other nodes under contrib (Design, Regression, Signal-Processing)

Nodes under core (chemometrics and microarray)

Edits

Dissection

The edit node is capable of editing and transforming data in SciCraft. Table 3.18 gives a summary of the EditNode. A detailed description of this node is provided in section 3.3.

Table 3.17: Summary of the dissection.
Name Dissection
Category Tools
Description Node that can remove parts of a dataset.
Usage Can alter the data you use for calculations.
Parameters None
Input ports Matrix
Output ports Matrix


Edit node

The edit node is capable of editing and transforming data in SciCraft. Table 3.18 gives a summary of the EditNode. A detailed description of this node is provided in section 3.3.

Table 3.18: Summary of the EditNode.
Name EditNode
Category Edit
Description Node that can remove one or several rows from a dataset.
Usage An edit node makes use of a command string to edit the data it receives. Its intention in later editions of SciCraft is to make it possible to select data in a plot node and then tell the edit node to remove these elements.
Parameters Command
Input ports commandInput - the enter point for data.
Output ports An exit port for the edited data.


Gaussdal node

This node is yet highly experimental. Use it at your own risk, and do not expect it to behave as intended. It has basic functionality, but lacks stability and extended functionality. The Gaussdal node takes as input molecule data from a GaussDal database query, and maps the molecules relative to a reference molecule. The node is connected to Octave's Procrustes routine, running this routine on the mapped molecules. A summary of the node is found in table 3.19. A more detailed description can be found in section 3.4.


Table 3.19: Summary of the Gaussdal node.
Name Gaussdal node
Category Edit
Description Molecule mapper for Gaussdal database data.
Parameters Mapping parameters
Input ports Data from Input-node reading Gaussdal files, column names, optional molecule mapping matrix.
Output ports Molecule data, mapping (not implemented yet)


MoleculeEditor

The MoleculeEditor node is capable of showing and editing mulecules in a 3D window. Table 3.20 gives a summary of the MoleculeEditor node.


Table 3.20: Summary of the MoleculeEditor.
Name MoleculeEditor
Category Tools
Description Node that can show and edit molecules in 3D.
Usage Node that can show and edit molecules in 3D.
Parameters None
Input ports Molecule data.
Output ports Molecule data.


View

Nodes capable of inspecting data currently in the nodes after running the diagrams. These nodes are summed up in (table 3.21 and 3.22).


Table 3.21: Summary of the ScreenDump node.
Name ScreenDump
Category View
Description Node that is used to make plot based on calculated data in the Module Diagram.
Usage This node's task is to save screendumps to file.
Parameters None
Input ports None
Output ports Imagefile


SpreadSheet Node

The spreadsheet node may take several matrices, vectors and scalars on its input port and display them in a spreadsheet. Table 3.22 gives a summary of the SpreadSheet Node, while a more detailed description is presented in section 3.5.


Table 3.22: Summary of the SpreadSheet Node.
Name SpreadSheet Node
Category View
Description Node that displays the selected data in a familiar spreadsheet.
Usage View matrices, vectors and scalars in a spreadsheet representation. It is possible to save all the data that is in the spreadsheet to file, and to make selections which may be copied to the clipboard, for pasting it into another application.
Parameters None
Input ports matrices - the entry point for data.
Output ports None




SciCraft Development Team