Data Mining & Exploration
We make science discovery happen
Clustering SOM help page
This is the page dedicated to help users of data mining web application (the beta release currently available here), in case of selection of SOM (Clustering with SOM) Model to make experiments. This page is also directly reachable from the web application, in case users select the help button. The following contents are mainly dedicated to assist the user during the model parameter selection and setup, by giving details about each parameter, its role in the model, default value and suggestions about the right choice.
There are 3 sub-sections, related with the use cases that users can select to perform experiments, respectively Train, Test and Run.
Functionality: Clustering with SOM model
Parameter specifications
Use Case: TRAIN
- Input file
- Dataset Type
- 0: used for .txt and .dat types
- 1: used for .csv, .votable. and .fits (fits-table) types
- 2: used for .jpg, .jpeg, .png and .gif types
- 3: used for .fits (fits-image) type
- Configuration file
- Input nodes
- Output rows
- Output columns
- Output dimension
- 0: 2D grid
- 1: 3D grid
- Normalize data
- 0: no normalized
- 1: yes normalized
- Initial learning rate
- Neighbor size
- Epochs
- Final learning rate
this parameter is a required field!
Input dataset file.
this parameter is a required field!
Integer indicating the type of input data format
Accepted entries are:
To be used only in case of resume training (from a previous training session)
this parameter is a required field!
Number of neurons (first layer), (always integer greater than 0).
T number of neurons of this layer corresponds to input features of the dataset
this parameter is a required field!
Number of rows of output layer (Kohonen layer), always greater than 1.
this parameter is a required field!
Number of columns of output layer (Kohonen layer), always greater than 1.
type of output layer dimension
If left empty, its default is 0 (2D)
Choice to force the normalization of input data in [-1, +1]
If left empty, its default is 0 (i.e. no normalized data)
Initial learning rate, Real number in ]0, 1[
If left empty, its default is 0.7
Initial neighbor size within the neurons of output grid. Integer not less than 0.
If left empty, its default is 3
Number of training epochs. Integer number greater than 0.
If left empty, its default is 200
final learning rate, Real number in ]0, initial learning rate[
If left empty, its default is 0.005
(See the user manual for more details)
Functionality: Clustering with SOM model
Parameter specifications
Use Case: TEST
- Input file
- Dataset Type
- 0: used for .txt and .dat types
- 1: used for .csv, .votable. and .fits (fits-table) types
- 2: used for .jpg, .jpeg, .png and .gif types
- 3: used for .fits (fits-image) type
- Dataset Target file
- Configuration file
this parameter is a required field!
Input dataset file.
this parameter is a required field!
Integer indicating the type of input data format
Accepted entries are:
this parameter is a required field!
target label file realted to the input dataset
this parameter is a required field!
Internal parameter setup after a training session
(See the user manual for more details)
Functionality: Clustering with SOM model
Parameter specifications
Use Case: RUN
- Input file
- Dataset Type
- 0: used for .txt and .dat types
- 1: used for .csv, .votable. and .fits (fits-table) types
- 2: used for .jpg, .jpeg, .png and .gif types
- 3: used for .fits (fits-image) type
- Configuration file
this parameter is a required field!
Input dataset file.
this parameter is a required field!
Integer indicating the type of input data format
Accepted entries are:
this parameter is a required field!
Internal parameter setup after a training session
(See the user manual for more details)
Copyright © 2010 DAME Collaboration. All right reserved.