A super powerful generic classifier that allows user to upload dataset, do EDA and print all summary statistics by a button-click. User selects one of 13 algorithms and hyperparameters using UI and train and evaluate model in one click. Bonus: User can plot 7 plots in one click. The application can handle .CSV and .XLSX files.
The 13 classifiers allowed are RANDOM FOREST , SVC, KNN, XGBOOST, ADABOOST, HISTGRADIENTBOOSTING, LOGISTIC REGRESSION,DECISION TREE, GRADIENT BOOSTING, LIGHTGBM, GAUSSIAN NAIVE BAYES, BERNOULLI NAIVE BAYES, NEURAL NETWORK. Once user chooses a classifier, corresponding hyper-parameters are automatically populated and once user selects his values, the model is trained and evaluation metrics are displayed.
Finally the user can choose a particular plot from 7 choices and the plot is created by a single button click with helpful information about how to use that plot for interpretation and decsion-making.
The salient features of the application are: (1) It can be deployed as a stand alone application on any operating system. (2) It can be deployed as a web application on cloud. (3) It can handle multi-class classification and binary classification both cases. (4) Basic data pre-processing steps like null-imputation, scaling and encoding categotical variables are built-in the model-training pipeline. So, no need of 100% clean data. (5) In a single window and by using just a few controls user can view summary statistics, all evaluation metrics and complete classification report and many types of plots.