Linear Regression, Logistic Regression, ML Pipeline

Предиктивный анализ оттока клиентов

Develop a model to predict which retail customers will respond to a marketing campaign. Logistic Regression shows the best performance.

A Comprehensive Guide to Titanic Machine Learning from Disaster

Light-weight package for classification metrics computed on streams or minibatches of data. Mainly for area under the curve (AUC) of precision-recall (PR) or receiver operating characteristic (ROC) curves. Supports multi-class setting with either macro- or micro aggregation..

Identify which customer is willing to possess the insurance policy, so we campaign efficiently.

The company has collected historical customer and claims data and wants to use it to develop a machine learning model that can predict whether a customer will file an insurance claim in the next year.

Results of binary classification of Yelp reviews as pertaining to conventional or alternative medicine using random forests

Build repository for brambox - https://gitlab.com/eavise/brambox

Predicting Bank Term Deposit Subscribers using Decision Trees

Repository where it is intended to address the challenge of LATAM Airlines, which consists of predicting the probability of delay of flights that land or take off from the Santiago de Chile airport (SCL)

98% accurate - This stroke risk prediction Machine Learning model utilises ensemble machine learning (Random Forest, Gradient Boosting, XBoost) combined via voting classifier. We tune parameters with Stratified K-Fold Cross Validation, ROC-AUC, Precision-Recall Curves and feature importance analysis.

This notebook describes how to compute and derive insights from various classification evaluation metrics.

Run histogram-based gradient boosted trees binary classifier on generated data and interpret results with standard metrics, SHAP, and supervised clustering

Resampling exercise to predict accuracy, precision, and sensitivity in credit-loan risk

A Portuguese hotel group seeks to understand reasons for its excessive cancellation rates.

The project involves using machine learning techniques, like RandomForestClassifier and MLP, to predict whether a song will be popular or not based on its acoustic features. The input consists of various acoustic and metadata features, while the output is a binary classification.

Sampling unbalanced dataset using SMOTE and creating a classifier to classify if a HR will stay or leave.