<< Click to Display Table of Contents >> Current:  Advanced Analytics > Algorithm  Logistic Regression

Logistic regression is a supervised problem of machine learning algorithms, which mainly solves two classification problems;

 

How-To-Use:

The input of this node is a data node. After setting the logistic regression, you can view the performance indicators by connecting the table view; connecting the picture view to view the ROC curve.

 

Precautions:

Logistic regression is more sensitive to outliers, and needs to be converted for nonlinear features; continuous values need to be discretized (in the case of inconsistent dimensions).

 

❖Configuration

After adding the Logistic regression node to the experiment, you can set the Logistic regression node through the "Configuration" page on the right.

[Proportion of training set] Usually, the proportion of the training set in the entire data set is 0.8, and the rest is used as the verification set.

[Positive example label] Please fill in None or the positive example label (integer or text type) in the target column. When filling in None, the default 1 is a positive example, and there are only (0,1) or (-1,1) in the target column .

[Optimizer] The optimizer is used to optimize system parameters; the corresponding optimization algorithms are as follows:'newton' refers to Newton-Raphson;'bfgs' refers to Broyden-Fletcher-Goldfarb-Shanno (BFGS);'lbfgs' refers to limited -memory BFGS;'powell' refers to the improved version of "Powell" method;'cg' refers to conjugate gradient;'ncg' refers to Newton-conjugate gradient;'basinhopping' refers to basin-hopping.

[Independent variable] Select the feature column of the front node, which can be multiple columns.

[Dependent variable] Select the target column of the pre-node, there can only be 1 column, data with categorical attributes.

Right-click menu of logistic regression:

❖Run Logistic Regression Node

Run the node, pass the data to DM-Engine for calculation, and get the output result.

 

❖Reset Logistic Regression Node

The node that has been running is reset, the returned result is deleted, and the node status is changed to not running.

 

❖Rename Logistic Regression Node

In the right-click menu of the Logistic Regression node, select "Rename" to rename the node.

 

❖Refresh Logistic Regression Node

In the right-click menu of the Logistic Regression node, select "Refresh" to update the synchronization data or parameter information.

 

❖Save as Composite Node

In the right-click menu of the Logistic Regression node, select "Save as Composite Node",The selected node can be saved as a composite node to realize a multiplexing node, and the parameters of the saved node are consistent with the original node.

 

❖Cut Logistic Regression Node

In the right-click menu of the Logistic Regression node, select "Cut" to realize node cutting operation.

 

❖Copy Logistic Regression Node

In the right-click menu of the Logistic Regression node, select "Copy" to realize node  replication operation.

 

❖Delete Logistic Regression Node

In the right-click menu of a Logistic Regression node, select "Delete" or click the delete key on the keyboard to delete the node and its input and output connections.