Documentation

1. Preparing Your Dataset

OptiAgent.ML processes structured tabular datasets in CSV format. Ensure your dataset has a clear header row. Categorical attributes will be auto-encoded using label encoders or one-hot vectors, and text columns will be converted using TF-IDF vectorization.

2. Running the AutoML Pipeline

Once you upload your CSV file and select the target label column:

• The agent audits the dataset for target leakage and drops proxy features.
• A series of classifiers/regressors (XGBoost, Random Forest, SVM) are executed.
• The leaderboards are computed and evaluated based on accuracy, F1-scores, or R2 metrics.

3. Model Exporting

After training completes, you can download the serialized best model (in PKL format) along with Python source templates to integrate the predictions directly into your production servers.