Introduction and Usage Guide for the xeasy-ml Machine Learning Framework

1. Introduction xeasy-ml is a Python machine‑learning framework that bundles common ML components and methods, allowing users to build models via configuration, automatically generate box plots, histograms, and perform feature‑correlation analysis. The project is hosted on GitHub.

2. Installation Required dependencies: scikit‑learn >=0.24.1, pandas >=0.24.2, numpy >=1.19.5, matplotlib >=3.3.4, pydotplus >=2.0.2, xgboost >=1.4.2. Install the package with pip install xeasy-ml .

3. How to use

Create a new Python file pro_init.py and initialize the project:

from xeasy_ml.project_init.create_new_demo import create_project
import os
pro_path = os.getcwd()
create_project(pro_path)

The generated directory structure looks like:

├── Your_project
    ...
    ├── pro_init.py
    ├── project
    │   └── your_project

Run the project:

cd project/your_project
python __main__.py

Result files are stored under project/your_project_name/result/v1 , including box plots, histograms, model files ( demo.m ), prediction results, and evaluation scores.

4. Online usage manual After training, three files are needed for prediction: demo.m , log.conf , and feature_enginnering.conf . Use xeasy_ml.predict() or instantiate PredictionML :

self.ml = xml.prediction_ml.PredictionML(config='your ml.conf path', xeasy_log_path=xeasy_log_path)

For XGBClassifier only, you can omit the model config and load the model directly:

self.ml = xml.prediction_ml.PredictionML(xeasy_log_path=xeasy_log_path)
self.ml._model = XGBClassifier()
self.ml._model.load_model(model_path)

Before predicting, process test data:

self.ml._feature_processor.test_data = data_frame
self.ml._feature_processor.execute()
test_feature = self.ml._feature_processor.test_data_feature.astype("float64", errors='ignore')
predict_res = self.ml._model.predict(test_feature)