Madison AI - A project to make a graphical user interface for training and evaluating machine learning models without writing any code. The software is useful for doing basic tasks in machine learning.

Features -

Data preprocessing

Feature Scaling 

     1) Normalization, 2) Standardization

Feature Encoding

     1) One-Hot Encoding, 2) Label Encoding

Handling Missing values

Feature Selection 

     1) Variancethreshold, 2) SelectkBest, 3) Recursive Feature Elimination using cross validation

Hyperparameters optimization using GridSearchCV

Training estimators available

Classifiers - 25 Classifiers

Regressors - 40 Regressors

Evaluation metrics

1)   Classification - F1 score, Log loss, Accuracy, Average precision, AUC

2)  Regression - Mean squared error, Mean absolute error, Explained variance score, Max error, Mean squared log error, R2 score

Graphical Representation - ROC-AUC curve, Precision Recall curve, Bar Graph, Line Graph, Histogram, Scatter plot.

Estimators, metrics and some other functions were taken from scikit-learn library,  Pandas library is used for handling the datasets, Matplotlib library is used for ploting graphs, Pandastable is used to display the dataset. Trained models are saved as “.pkl” files.