Node generation via recursion

How can I create nodes recursively to generate tree in MATLAB

[KSSV]

https://in.mathworks.com/matlabcentral/answers/1633-how-to-make-a-tree

[Ashish Kumar]

Is there any way to create nodes explicitly?

Actually I want to generate the tree using recursion with the help of nodes.

Answer 1

[Shantanu Dixit]

Hi Ashish,

If I understood the query, there are particularly two context where recursively generating nodes can apply.

1. GUI based: Building a UI-based tree through recursion (using uitreenode)
2. Decision Tree Splitting: Recursively splitting the nodes based on the best feature from the data

GUI based: If you're working with hierarchical data (like a folder structure or a nested JSON), you can use recursion to dynamically add child nodes to a tree in the UI.

Here's a general idea of how you can approach this using 'uitree' and 'uitreenode' for a nested structure:

%{
dataStruct
├── A
│   ├── A1
│   └── A2
└── B
    ├── B1
    └── B2
%}

dataStruct = struct('A', struct('A1', [], 'A2', []), 'B', struct('B1', [], 'B2', []));

% Create the root UI tree
t = uitree('Parent', uifigure);
root = uitreenode(t,'Text','root');

% Recursive function to add nodes
function addNodes(parentNode, s)
    fields = fieldnames(s);
    for i = 1:numel(fields)
        child = uitreenode(parentNode, 'Text', fields{i});
        if isstruct(s.(fields{i}))
            addNodes(child, s.(fields{i}));  % Recursive call
        end
    end
end

addNodes(root, dataStruct);

Decision Tree Splitting: In MATLAB the recursive tree splitting for decision tree's is usually abstracted through the method calls like 'tree = fitctree(X, Y)'. However if you want to manually implement the decision tree splitting criteria you can:

1. Choose the best feature to split on (eg. Gini index or entropy)
2. Recur on the data to create left and right branches (child nodes)
3. Stop when a stopping condition is met (pure node / max depth)

Additionally you can refer to the following documentation of App building in MATLAB for more information:

uitree: https://www.mathworks.com/help/matlab/ref/uitree.html

uitreenode: https://www.mathworks.com/help/matlab/ref/uitreenode.html

Hope this helps!