.. _auto_ml_decision_tree_train:

*TRAIN OR CLASSIFY USING DECISION TREE*
=======================================

Trains a Decision Tree model on a selected dataset column or runs classification using a previously trained tree. Use this worker when interpretable, rule-based classification or regression is needed on tabular data stored in a d3VIEW database.

When to use
-----------

Tagged: ``classification``, ``decision_tree``, ``ml``, ``regression``, ``sklearn``, ``supervised_learning``, ``train``.

Inputs
------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20 20 20

   * - Label
     - ID
     - Type
     - Default
     - Required
     - Description
   * - Database
     - database_id
     - text
     - —
     - ✓
     - d3VIEW database reference containing the training dataset; select the target database from the dropdown — required to resolve the feature and label columns.
   * - Column Name
     - train_column
     - file
     - —
     - ✓
     - Name of the target (label) column within the selected database that the Decision Tree will learn to predict; must match the column header exactly.
   * - Drop Columns
     - drop_columns
     - scalar
     - 1
     - 
     - Comma-separated list of column names to exclude from training features; default value of 1 means no columns are dropped — override when uninformative or leaky columns need removal.

Outputs
-------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20

   * - Label
     - ID
     - Type
     - Description
   * - Train Model
     - train_model
     - text
     - Serialized Decision Tree model artifact (e.g., pickle/joblib file path or model reference) produced after training; consumed by downstream scoring or evaluation workers.

Disciplines
-----------

- ai_ml.model_selection
- ai_ml.supervised.classification
- ai_ml.supervised.regression

.. raw:: html

   <hr style="margin-top:2em">
   <p style="font-size:11px;color:#888">
   Auto-generated from <code>platform</code> schema. Worker id: <code>ml_decision_tree_train</code>. Schema hash: <code>6baa720c5963</code>. Hand-curated docs in <code>workerexamples/</code> override this page when present.
   </p>