Machine Learning with Python

所在平台: Coursera

课程主页: https://www.coursera.org/learn/machine-learning-with-python

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：用Python进行机器学习 课程概述：本课程深入探讨机器学习的基础知识，使用易于掌握且广泛应用的编程语言Python进行学习。课程将重点关注两个主要部分：首先，您将了解机器学习的目的以及其在现实世界中的应用；其次，您将获得机器学习主题的一般概览，包括监督学习与非监督学习、模型评估和机器学习算法。 通过实际案例的练习，您将看到机器学习如何在意想不到的方式中影响社会！ 在接下来的几周里，每周投入几个小时，您将获得： 1) 新增到简历的技能，例如回归、分类、聚类、Sci-kit Learn 和 SciPy； 2) 可以加入您作品集的新项目，包括癌症检测、经济趋势预测、客户流失预测、推荐引擎等； 3) 一份机器学习证书，以证明您的能力，可以在任何您喜欢的线上或线下地方分享，如LinkedIn和社交媒体。 如您选择参加本课程并获得Coursera课程证书，成功完成后，您还将获得IBM数字徽章。 课程大纲： 1. **机器学习简介**：探索基础机器学习概念，为使用Python进行实践建模做好准备。 2. **线性与逻辑回归**：研究机器学习中的两种重要回归技术——线性回归和逻辑回归，包括实际数据使用。 3. **建立监督学习模型**：构建并评估各种监督机器学习模型，用于解决分类与回归问题。 4. **建立非监督学习模型**：学习如何通过非监督学习技术揭示数据中的隐藏模式，应用K-Means等进行客户细分。 5. **评估与验证机器学习模型**：学习如何使用行业标准技术评估机器学习模型的有效性，解释模型评估指标。 6. **最终项目与考试**：应用和展示您在课程期间获得的全部技能，通过实际项目进行评估，并完成最终考试。 本课程为您提供了全面的机器学习知识体系，使您能够在这一领域中绽放光彩。

课程大纲

Name:Introduction to Machine Learning

Description:In this module, you will explore foundational machine learning concepts that prepare you for hands-on modeling with Python. You will explain the relevance of Python and scikit-learn in machine learning, summarize the IBM AI Engineering certification path, and classify common types of learning algorithms. You’ll outline the stages of the machine learning model lifecycle and describe what a typical day looks like for a machine learning engineer. You will also compare key roles in the AI field, identify widely used open-source tools, and learn to utilize scikit-learn to build and evaluate simple models.

Name:Linear and Logistic Regression

Description:In this module, you will explore two essential regression techniques used in machine learning—linear and logistic regression. You’ll explain the role of regression in predicting outcomes, describe the differences between simple and multiple linear regression, and apply both using scikit-learn on real-world data. You will also interpret how polynomial and non-linear regression models capture complex patterns. The module introduces logistic regression as a classification method and guides you in training and testing classification models effectively. To support your learning, you’ll receive a Cheat Sheet: Linear and Logistic Regression that summarizes key concepts, formulas, and use cases.

Name:Building Supervised Learning Models

Description:In this module, you will build and evaluate a range of supervised machine learning models to solve both classification and regression problems. You’ll start by describing how classification models predict categorical outcomes, and implement multi-class classification strategies using real-world data. You’ll then explore how decision trees make predictions and apply them to both classification and regression tasks. The module also covers using support vector machines (SVM) for fraud detection, applying K-Nearest Neighbors (KNN) for customer classification, and training ensemble models like Random Forest and XGBoost to improve accuracy and efficiency. You’ll differentiate bias and variance in model performance and explore how ensemble methods help balance this tradeoff. To support your learning, you’ll receive a Cheat Sheet: Building Supervised Learning Models with key terms, model types, and evaluation tips.

Name: Building Unsupervised Learning Models

Description:In this module, you will learn how unsupervised learning techniques uncover hidden patterns in data without using labeled responses. You’ll describe clustering concepts and apply K-Means to real-world customer segmentation tasks. You’ll also compare DBSCAN and HDBSCAN models to identify dense clusters in spatial data. Moving beyond clustering, you’ll explore dimensionality reduction as a tool for simplifying high-dimensional datasets. You’ll apply PCA to uncover key components and use advanced techniques like t-SNE and UMAP to visualize data structure. To support your learning, you’ll receive a Cheat Sheet: Building Unsupervised Learning Models, highlighting core methods, practical use cases, and comparison guidelines.

Name:Evaluating and Validating Machine Learning Models

Description:In this module, you will learn how to assess the effectiveness of machine learning models using industry-standard evaluation and validation techniques. You’ll explain key classification and regression metrics, evaluate models using real-world data, and interpret results with tools like confusion matrices and feature importance charts. You'll explore how to assess clustering quality in unsupervised learning and apply cross-validation to reduce overfitting. The module also introduces regularization methods to improve model generalization and reduce feature complexity. Finally, you'll build complete machine learning pipelines and optimize them with GridSearchCV, while identifying common pitfalls like data leakage. To support your learning, you’ll receive a Cheat Sheet: Evaluating and Validating Machine Learning Models covering key metrics, techniques, and model tuning strategies.

Name:Final Project and Exam

Description:In this final module, you will apply and demonstrate the full range of skills you have gained throughout the course. You will start with a practice project using the Titanic dataset to build and optimize classification models using pipelines, cross-validation, and hyperparameter tuning. Then, you will complete the final project by developing a rainfall prediction classifier using historical weather data. This includes data cleaning, feature engineering, model building, and evaluating performance. To conclude the course, you will take a graded final exam that tests your knowledge across all six modules. This module gives you the opportunity to showcase your learning in both practical and theoretical contexts.

课程评论(0条)