发帖
雷达卡
R is an open source programming language with a lot of facilities for problem solving through statistical computing. At the time of writing this, there are more than 5K packages available in CRAN repository. Below are few reasons that make R language stand out and why you should really consider learning the language.
R is a language and an environment for everything related to data. But what is ‘everything’?.
It includes statistical computing, data mining, data analysis, machine learning, predictive modelling, quantitative analysis, optimisation and operations research etc – all of which are somewhat inter-related terms.
Who uses all this ?
Data scientists, analysts, statisticians, quantitative analysts, forecasters, bio-statisticians, financial analysts, research scientists. These are some of the professions where R is commonly used. But, is R limited to these guys? NO, and not necessary!
Here are some use cases..
If you are involved with anything directly or indirectly related to math, you should consider to learn and use R.
1. May be you want to predict the number of people expected to visit your shop over the next week, or, what is the right combination of meal combo you should put up for sale so it is of liking for the maximum portion of your customers, or, what plan or product should you offer your customers depending on what they already purchased so there is more chance he/she will take your offering?
2. You are a supply chain fellow who wants to know how much of inventory to stock of each part in which of your warehouses, or perhaps when should you exactly make an order for a part (because it can take a month or so for the new batch to reach your storehouse while your customers are continuing to empty your shelves at an irregular pace).
3. You are a bio-statistician, interested to find out if women over 45 years of age and living in high altitude, are more susceptible to heart related diseases?
Getting it? Well, that’s just the surface! There are innumerable variety of problems that can be unearthed and solved in almost every field // business.
R is not just a software or a programming language or a excellent visualization tool, where you can write algorithms to solve problems. It is backed by the works of a community of statisticians, scientists, and engineers in the form of packages that are freely available for you and everyone to use. The bad news is, you still need to write code. But the great news is the rich collection of packages does the heavy lifting for you, that your options have been widened and efforts to arrive at the solution is minimized. In fact, there is more packages getting created every other day. I think it is ok to say, whenever a new technique/ML method/algo is invented, it almost always, first shows its face in R before being implemented in other platforms. This is mainly because R is the medium de facto used by the researchers, professionals and educational institutions alike.
So what are the advantages of using R ?
1. Unlimited Possibilities!R is a door to a whole world of problem solving and research through itsapplications in a number of domains. You may be into quantitative finance, biologist or a supply chain specialist or anything under the cloud, there is a ton of powerful things you can do with R with much less effort. Currently, there are more than 6.5K packages addressing problems in a vast variety of domains.
The current version is a result of collaborative effort of programmers, researchers and contributors from all over the world. It gets better by the day.
What does this mean to you? You can install R and will have access to all future releases, updates and the powerful packages at CRAN always available to you free of cost.
The creators of R has laid down a structured approach into the documentation procedure right at the beginning. So it is a lot more easier to get around compared to other open-source alternatives.
More educational institutions, researchers and authors have adopted R as their primary medium of work. This means the latest techniques, books and research papers publish their findings in R before getting implemented in other software.
Compared to other popular alternatives, R throws lesser errors, especially with respect to data formatting, etc. This is because R takes of many things in the background. Comparatively, R users spend much less time on debugging and data formatting and rather focus more on problem solving approaches and solutions. It gives more control to the user. You will soon be super-fast and efficient as most tenured R programmers are. You will probably appreciate this more when you get your hands wet.
However, R is said to have a steep learning curve. As with learning any programming language, you will have to crawl before you walk, but the fruits of your efforts will be multiple folds compared to the ‘pain’ of the learning exercise. R is a preferred language for data science and problem solving. A number of corporations are setting up their own in-house data science factory which has created a huge demand for knowledgeable data science workers.
Because of its potential to create enormous value, R programming is arguablythe the highest paid skill as of 2015!
7月24-28日“机器学习及R应用”集中短训现场班
授课方式：思想原理 + 数学精髓 + R经典案例
讲师介绍：
本课程由山东大学经济学院陈强教授亲授。陈强教授获得北京大学经济学学士、硕士，美国Northern Illinois University数学硕士、经济学博士，现为数量经济学博士生导师，在统计学、计量经济学及机器学习领域具有深厚的功底，2010年入选教育部新世纪优秀人才支持计划。陈强老师著有畅销研究生教材《高级计量经济学及Stata应用》（第2版，高教社，2014），并特别擅长深入浅出、直指人心地介绍数据分析原理，深受广大学生们的喜爱，其现场班常常人满为患、好评如潮。
开课信息：
时间：2021年7月24-28日(五天)
地点：北京市海淀区
费用：5200元/ 4500元(本科及硕士在读优惠价)；食宿自理
安排：上午9:00-12:00；下午2:00-5:00；答疑
报名：http://www.peixun.net/main.php?mod=buy&cid=1436
培训目的和特色：
机器学习早期为人工智能的分支，后来也有不少统计学家加入，最近一、二十年因为其预测精度迅速提高而走红，并在业界有着广泛的应用。可以预见，在未来三十年，几乎所有行业都会因机器学习的深刻冲击而改变。MIT名誉校长Eric Grimson曾预言，机器学习会成为像Word一样的工具。而谁先掌握此工具，则可占得先机，成为时代的弄潮儿（至少不会落伍）。
基于机器学习的通用性，本次“机器学习及R应用”五天现场班将面向所有行业与学科的人士、老师与学生（包含经管社科、医学卫生等领域）。
本课程的最大特色在于“一站式服务”，从机器学习的原理、数学推导，到R语言命令与经典案例，无不精心设计、丝丝入扣，理论联系实操，让学员们迅速理解机器学习的精髓，并掌握最为流行的数据科学软件R语言操作。
陈强老师将从零开始，介绍R语言的精华，让你迅速上手！
Why R?
√ R是统计学家发明的专门用于统计计算的语言
√ R是统计学家的母语
√ R中的统计“包”(package)最多，且增长迅速
√ 统计学顶级期刊的新发表论文一般带有相应的R包
√ R是免费开源的，在学界与业界均有很多用户
培训内容目录:
第1讲 机器学习引论
(1) 什么是机器学习
(2) 机器学习的分类与术语
(3) 案例：垃圾邮件过滤；手写体数字识别；图像识别；自动驾驶
第2讲 R语言快速入门
(1) Why R？
(2) 安装R与RStudio
(3) R的对象(vector, matrix, data frame,list)
(4) 面向对象的函数式语言
(5) R语言画图
第3讲 数学回顾
(1) 梯度向量
(2) 方向导数
(3) 梯度下降
(4) 向量微分
(5) 最优化
第4讲 线性回归
(1) OLS
(2) 过拟合与泛化能力
(3) 偏差与方差的权衡
(4) 交叉验证
(5) R案例：多项式回归的过拟合；波士顿房价
第5讲 逻辑 回归
(1) Logit
(2) 几率比
(3) 灵敏度与特异度
(4) ROC与AUC
(5) 科恩的kappa
(6) R案例：泰坦尼克号旅客的存活
第6讲 多项逻辑 回归
(1) 多项Logit
(2) R案例：识别玻璃类别
第7讲 判别分析
(1) 线性判别分析(LinearDiscriminant Analysis)
(2) 二次判别分析(QuadraticDiscriminant Analysis)
(3) 费雪判别分析(FisherDiscriminant Analysis)
(4) R案例：鸢尾花品种的归类
第8讲 朴素贝叶斯
(1) 朴素贝叶斯(Naive Bayes)
(2) 拉普拉斯修正(LaplacianCorrection)
(3) R案例：垃圾邮件的识别
第9讲 惩罚回归
(1) 高维回归的挑战
(2) 岭回归(Ridge Regression)
(3) 套索估计(Lasso)
(4) 弹性网估计(Elastic Net)
(5) R案例：前列腺癌的影响因素
第10讲 K近邻法
(1) 回归问题的K近邻法
(2) 分类问题的K近邻法
(3) R案例：摩托车撞击实验数据；模拟混合数据；威斯康辛乳腺癌的诊断
第11讲 决策树
(1) 分类树(Classification Tree)
(2) 分裂准则(错分率、基尼指数、信息熵)
(3) 成本复杂性修枝
(4) 回归树(Regression Tree)
(5) R案例：波士顿房价；葡萄牙银行市场营销
第12讲 随机森林
(1) 集成学习(Ensemble Learning)
(2) 装袋法(Bagging)
(3) 随机森林(Random Forest)
(4) 变量重要性(Variable Importance)
(5) 偏依赖图(Partial Dependence Plot)
(6) R案例：波士顿房价；声呐信号的分类
第13讲 提升法
(1) 自适应提升法 (AdaBoost)
(2) AdaBoost的统计解释
(3) 梯度提升法 (Gradient Boosting Machine)
(4) XGBoost
(5) R案例：波士顿房价；过滤垃圾邮件；识别玻璃类别
第14讲 支持向量机
(1) 最大间隔分类器(MaximalMargin Classifier)
(2) 软间隔分类器(Soft MarginClassifier)
(3) 支持向量机(Support Vector Machine)
(4) 核技巧(Kernel Trick)
(5) 支持向量回归(SupportVector Regression)
(6) R案例：模拟数据；过滤垃圾邮件；识别手写数字；波士顿房价
第15讲 人工神经网络
(1) 人工神经网络的思想
(2) 感知机（Perceptron）
(3)前馈神经网络(Feedforward Neural Network)
(4) 激活函数(Activation Function)
(5) 反向传播算法(Back-propagation Algorithm)
(6) 随机梯度下降(Stochastic Gradient Descent)
(7) 神经网络的过拟合与正则化
(8) 卷积神经网络(Convolution Neural Network)
(9) 深度学习的发展
(10) R案例：波士顿房价；声呐信号的分类；鸢尾花品种的分类
第16讲 非监督学习之主成分分析
(1) 总体中的主成分分析
(2) 样本中的主成分分析
(3) 方差分解与降维
(4) 主成分回归(PrincipalComponent Regression)
(5) R案例：左右耳听力；香港回归的经济效应
第17讲 非监督学习之聚类分析
(1) K-均值聚类(K-meansClustering)
(2) 分层聚类(Hierarchical Clustering)
(3) 树状图
(4) 基于相关系数的距离
(5) R案例：模拟数据；鸢尾花品种的归类
第18讲 数据科学的R语言
(1) 何为数据科学
(2) 管道算子(Pipe Operator)
(3) R包tidyverse(输入数据、数据清理、数据变换)
(4) R包ggplot2(高阶画图)
(5) R包caret(机器学习的统一接口)
(6) R案例：R包tidyverse的自带案例；威斯康辛乳腺癌的诊断
第19讲（Bonus Lecture） 机器学习在经管社科的应用
精读几篇在经管社科顶刊发表的经典机器学习论文
不难看出，本次课程可谓干货满满、奇货可居。更难得可贵的是，主讲老师陈强教授具有丰富的教学经验、激情与魅力，是广大计量学子心目中真正的“计量男神”，尤其擅长化繁为简、直指人心，让学员们迅速上手新知识与技能。
跟着陈强老师，五天入门机器学习，登堂入室，立竿见影，赶上时代的步伐！
优惠：
现场班老学员9折优惠；
同一单位三人以上同时报名9折优惠；
同一单位六人以上同时报名8折优惠；
以上优惠不叠加。
报名流程：
1：点击“http://www.peixun.net/main.php?mod=buy&cid=1436”，网上填写信息提交；
2：给予反馈，确认报名信息；
3：网上订单缴费（需要刷卡或对公转账的请报名后与我们联系）；
4：开课前一周发送课程电子版讲义，软件准备及交通住宿指南。
联系方式：
尹老师
电话: 010-53352991
QQ: 42884447
邮箱: yinna@pinggu.org
微信：yinyinan888
提升卡
置顶卡
沉默卡
变色卡
抢沙发
千斤顶
显身卡
京公网安备 11010802022788号
