|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/fundamentals-of-machine-learning-hindi/
课程评论：没有评论
课程名称：全方位：机器学习、深度学习、自然语言处理、AWS 部署 [印地语][Python] 课程概述： 本课程旨在全面覆盖机器学习的各个概念，适合任何人学习，无需先前的机器学习基础。课程内容包括自然语言处理和深度学习的基础知识。具体涵盖的主题如下： 章节一：机器学习入门 - 机器学习概述及类型 章节二：环境设置 - 安装Anaconda，使用Spyder和Jupyter Notebook，安装相关库 章节三：云环境创建（AWS） - 创建EC2实例，连接EC2，安装库，文件传输及执行Python脚本 章节四：数据预处理 - 处理空值、特征相关性检查、数据重塑、缺失值填补、特征缩放、标签编码和独热编码 章节五：监督学习-回归 - 简单线性回归、最小化成本函数（普通最小二乘法、梯度下降）、线性回归假设、虚拟变量、多元线性回归、回归模型性能评估（R²）、多项式线性回归 章节六：监督学习-分类 - 逻辑回归、K-近邻、朴素贝叶斯、模型保存与加载、分类模型性能评估（混淆矩阵） 章节七：无监督学习-聚类 - 划分算法：K均值算法、随机初始化陷阱、肘部法则；分层聚类：聚合型、树状图；基于密度的聚类：DBSCAN；无监督聚类性能评估（轮廓指数） 章节八：无监督学习-关联规则 - Apriori算法和关联规则挖掘 章节九：使用Flask部署机器学习模型 - 理解工作流程，服务器端和客户端编码，AWS上设置Flask，发送请求及响应处理 章节十：非线性监督算法 - 决策树回归和分类，支持向量机（SVM）分类、核SVM、软边界、核技巧 章节十一：自然语言处理 - 文本预处理技术（包括标记化、停用词去除、N-grams、词义消歧等），以及案例研究（垃圾邮件过滤器） 章节十二：深度学习 - 人工神经网络、隐藏层、激活函数、前向与反向传播，使用感知器在Python中实现门 章节十三：正则化和回归 - 过拟合与欠拟合，偏差与方差，正则化（L1和L2损失函数），Lasso和Ridge回归 章节十四：降维 - 特征选择（前向与后向），特征提取（PCA、LDA） 章节十五：集成方法 - 袋装（随机森林回归与分类）和提升（梯度提升回归与分类） 此课程将为学员提供广泛的机器学习知识及实践经验，旨在帮助其在数据科学领域取得成功。
This course is designed to cover maximum concepts of machine learning a-z. Anyone can opt for this course. No prior understanding of machine learning is required.Bonus introductions include Natural Language Processing and Deep Learning.Below Topics are covered Chapter - Introduction to Machine Learning- Machine Learning?- Types of Machine LearningChapter - Setup Environment - Installing Anaconda, how to use Spyder and Jupiter Notebook- Installing LibrariesChapter - Creating Environment on cloud (AWS)- Creating EC2, connecting to EC2- Installing libraries, transferring files to EC2 instance, executing python scriptsChapter - Data Preprocessing- Null Values- Correlated Feature check- Data Molding- Imputing- Scaling- Label Encoder- On-Hot EncoderChapter - Supervised Learning: Regression- Simple Linear Regression- Minimizing Cost Function - Ordinary Least Square(OLS), Gradient Descent- Assumptions of Linear Regression, Dummy Variable- Multiple Linear Regression- Regression Model Performance - R-Square- Polynomial Linear RegressionChapter - Supervised Learning: Classification- Logistic Regression- K-Nearest Neighbours- Naive Bayes- Saving and Loading ML Models- Classification Model Performance - Confusion MatrixChapter: UnSupervised Learning: Clustering- Partitionaing Algorithm: K-Means Algorithm, Random Initialization Trap, Elbow Method- Hierarchical Clustering: Agglomerative, Dendogram- Density Based Clustering: DBSCAN- Measuring UnSupervised Clusters Performace - Silhouette IndexChapter: UnSupervised Learning: Association Rule- Apriori Algorthm- Association Rule MiningChapter: Deploy Machine Learning Model using Flask- Understanding the flow- Serverside and Clientside coding, Setup Flask on AWS, sending request and getting response back from flask serverChapter: Non-Linear Supervised Algorithm: Decision Tree and Support Vector Machines- Decision Tree Regression- Decision Tree Classification- Support Vector Machines(SVM) - Classification- Kernel SVM, Soft Margin, Kernel TrickChapter - Natural Language ProcessingBelow Text Preprocessing Techniques with python Code- Tokenization, Stop Words Removal, N-Grams, Stemming, Word Sense Disambiguation- Count Vectorizer, Tfidf Vectorizer. Hashing Vector- Case Study - Spam Filter Chapter - Deep Learning- Artificial Neural Networks, Hidden Layer, Activation function- Forward and Backward Propagation - Implementing Gate in python using perceptronChapter: Regularization, Lasso Regression, Ridge Regression- Overfitting, Underfitting- Bias, Variance- Regularization- L1 & L2 Loss Function - Lasso and Ridge RegressionChapter: Dimensionality Reduction- Feature Selection - Forward and Backward- Feature Extraction - PCA, LDAChapter: Ensemble Methods: Bagging and Boosting- Bagging - Random Forest (Regression and Classification)- Boosting - Gradient Boosting (Regression and Classification)
