Zero to Agile Data Science

所在平台: Udemy

课程主页: https://www.udemy.com/course/agile-data-science/

课程评论:没有评论

第一个写评论        关注课程

课程简介

课程名称:从零到敏捷数据科学 课程概述:本课程将教你如何将敏捷数据科学技术应用于分类问题,包含三个项目:预测信用卡欺诈、预测客户流失和预测财务危机。每个项目将分为五个迭代,从“第一天”到“第五天”,循序渐进地引导你从简单的随机森林分类器发展到经过调优的五个分类器组合(包括XGBoost、LightGBM、梯度提升决策树、额外树和随机森林),并在上采样数据上进行评估。该课程非常适合中级数据科学家,旨在扩展他们的技能,内容包括: - 自动检测原始数据中的无效列(第一天) - 为不平衡数据集创建自定义指标(第一天) - 四种数据重采样技术(第二天) - 处理缺失值(第二天) - 两种特征工程技术(第三天) - 四种特征降维技术(第三天) - 减少内存占用(第三天) - 在GridSearchCV中设置自定义评分函数(第四天) - 更改XGBoost的默认评分指标(第五天) - 构建元模型(第五天) 学员将获得完整的Jupyter笔记本、源代码及可复用函数库,以便在自己的项目中随时使用。

课程评论(0条)

课程详情

You will learn how to apply Agile Data Science techniques to Classification problems through 3 projects - Predicting Credit Card Fraud, Predicting Customer Churn and Predicting Financial Distress. Each project will have 5 iterations labelled ‘Day 1' to ‘Day 5' that will gently take you from a simple Random Forest Classifier to a tuned ensemble of 5 classifiers (XGBoost, LightGBM, Gradient Boosted Decision Trees, Extra Trees and Random Forest) evaluated on upsampled data.This course is ideal for intermediate Data Scientists looking to expand their skills with the following:Automated detection of bad columns in our raw data (Day 1)Creating your own metric for imbalanced datasets (Day 1)Four Data Resampling techniques (Day 2)Handling Nulls (Day 2)Two Feature Engineering techniques (Day 3)Four Feature Reduction techniques (Day 3)Memory footprint reduction (Day 3)Setting a custom scoring function inside the GridSearchCV (Day 4)Changing the default scoring metric for XGBoost (Day 5)Building meta-model (Day 5)Complete Jupyter notebooks with the source code and a library of reusable functions is given to the students to use in their own projects as needed!

课程标签

0人关注该课程

主题相关的课程