random forest

Documents à disposition des stagiaires pour la formation du 16 avril 2024.

Evaluación de distintos modelos para clasificar aguas entre potables y no potables

Generates a model predicting substrate class (rock, mixed, sand, or mud) using random forest and a stack of environmental predictors. Model is tested against withheld testing observations. Model can be run using a 1-step or 2-step approach. The 2-step approach generates a binary rock/not-rock layer and a mixed/sand/mud layer and the two are combined. This approach was created to reduce the prevalence of predicted rock in the output raster. Substrate index layers (densities per class) can also be produced.

Финальная работа по курсу Введение в DS от Skillbox. Цель: анализ данных сервиса СберАвтоподписка, предсказание совершения целевого действия пользователя и подготовка модели к деплою в продакшн. Метрика: ROC-AUC

project for module 33.Airflow in the course "Basic Data Science" at Skillbox.

Machine Learning - Binary Classification

Machine Learning - Multiclass Classification

Machine Learning - Multiclass Classification

Machine Learning - Classification Problem

Machine Learning - Multiclass Classification

Machine Learning - Regression Problem

Entity Embeddings of categorical variables implementation in TensorFlow 2.0, with a comparison in prediction performances on a Kaggle dataset, against One Hot Encoding using a Random Forest.

King County House Sales