This is a python package useful for the automated logging and visulaization of metrics for machine learning tasks
Project description
This project is about a python package for automated logging and visualization of metrics of classfication and regression algorithms in machine learning
Features
- The module covers both regression and classification tasks.
- The module integrates a wide range of metrics related to classification and regression
- The module can provide a barplot of the specified metrics from the specified subset of data
- The module can provide the confusion matrix for all the different ML algorithms
Functions
- Train and log for classification. This function applies around 12 different classification algorithms as mentioned below:- 'svm-linear' 'svm-rbf' 'svm-poly' 'knn' 'naive bayes' 'decision tree' 'random forest' 'adaboost' 'gradient boost' 'xgboost' 'logistic regression' 'bagging classifier'
All the classifiers are trained on the datasets and the results (accuracy, precision, recall, F1) are logged onto a dataframe which is displayed to the user.
-
Get confusion matrix. This function helps to get the confusion matrices for all the classification algorithms on the specified set (training/validation)
-
Display metric plots. This function plots a barplot for the comparative analysis of the classication algorithms on the specified metric and on the specified subset.
-
Train and log for regression. This function applies around 11 different regression algorithms as mentioned below:- 'linear regression' 'sgd regression' 'ridge regression' 'elastic net' 'decision tree regression' 'random forest regression' 'adaboost regression' 'gradient boost regression' 'xgboost regression' 'bagging regression' 'hist gradient boosting regression'
All the classifiers are trained on the datasets and the results (mae, mse, msle, median error, mape, max error) are logged onto a dataframe which is displayed to the user.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for AutoLogging_ML-0.0.2.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 2cfa20fc92289da4f776e104aff59fab8e63f140ce3e0049b5f8209d4bd9531a |
|
MD5 | ca6e95ef62fc0e7d136e8fe7f9f657de |
|
BLAKE2b-256 | 806bcab761be63705a733c5b21da7187ab16a3b609d756c5cb59ef1682ebea36 |