R Deep Learning Essentials

相似文件 换一批

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

R Deep Learning Essentials
               
               
            
                                            

R Deep Learning Essentials by Dr. Joshua F. Wiley
Mar. 30, 2016 | ISBN: 1785280589 | 170 Pages | AZW3/MOBI/EPUB/PDF (conv) | 10.04 MB

深度学习是机器学习的一个分支，H2O包已经在研究深度学习的数字科学家中越来越流行。它有着超级内存管理能力以及与多节点大数据平台的完全交互能力。

主要内容

使用R语言来根据深度学习概念，为无人管理的数据构建一个算法。
精通将面临的一般问题，如：拟合数据、异常数据集、图像识别和建模时的性能调整。
建立与模型相关的神经网络、预报和深度预报。

目录

Getting Started with Deep Learning
Training a Prediction Model
Preventing Overfitting
Identifying Anomalous Data
Training Deep Prediction Models
Tuning and Optimizing Models
Bibliography

R Deep Learning Essentials - Dr. Joshua F. Wiley.rar (10.04 MB, 需要: 15 个论坛币) 本附件包括：

• R Deep Learning Essentials - Dr. Joshua F. Wiley.mobi
• R Deep Learning Essentials - Dr. Joshua F. Wiley.pdf
• R Deep Learning Essentials - Dr. Joshua F. Wiley.azw3
• R Deep Learning Essentials - Dr. Joshua F. Wiley.epub

2016-4-26 11:45:47 上传

需要: 15 个论坛币  [购买]

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏14 回帖

关键词：Essentials Essential Learning earning Learn management learning machine memory become

谢谢分享

好书就是贵，先买了慢慢学习，多谢分享。

楼主快看一下，下载到一半时提示有病毒，怎么回事？

大家都木有问题。

1. source("checkpoint.R")
   
2. options(width = 70, digits = 2)
   
3. 
   
4. cl <- h2o.init(
   
5.   max_mem_size = "12G",
   
6.   nthreads = 4)
   
7. 
   
8. ## data setup
   
9. digits.train <- read.csv("train.csv")
   
10. digits.train$label <- factor(digits.train$label, levels = 0:9)
    
11. 
    
12. h2odigits <- as.h2o(
    
13.   digits.train,
    
14.   destination_frame = "h2odigits")
    
15. 
    
16. i <- 1:32000
    
17. h2odigits.train <- h2odigits[i, ]
    
18. 
    
19. itest <- 32001:42000
    
20. h2odigits.test <- h2odigits[itest, ]
    
21. xnames <- colnames(h2odigits.train)[-1]
    
22. 
    
23. 
    
24. 
    
25. system.time(ex1 <- h2o.deeplearning(
    
26.   x = xnames,
    
27.   y = "label",
    
28.   training_frame= h2odigits.train,
    
29.   validation_frame = h2odigits.test,
    
30.   activation = "RectifierWithDropout",
    
31.   hidden = c(100),
    
32.   epochs = 10,
    
33.   adaptive_rate = FALSE,
    
34.   rate = .001,
    
35.   input_dropout_ratio = 0,
    
36.   hidden_dropout_ratios = c(.2)
    
37. ))
    
38. 
    
39. 
    
40. system.time(ex2 <- h2o.deeplearning(
    
41.   x = xnames,
    
42.   y = "label",
    
43.   training_frame= h2odigits.train,
    
44.   validation_frame = h2odigits.test,
    
45.   activation = "RectifierWithDropout",
    
46.   hidden = c(100),
    
47.   epochs = 10,
    
48.   adaptive_rate = FALSE,
    
49.   rate = .01,
    
50.   input_dropout_ratio = 0,
    
51.   hidden_dropout_ratios = c(.2)
    
52. ))

复制代码

1. raining and predicting new data from a deep neural network
   
2. 
   
3. In this section we will learn how to train deep neural networks and use them to generate predictions on new data. The examples for this section will use the activity data we have worked with before, and the following code simply sets up the data:
   
4. 
   
5. use.train.x <- read.table("UCI HAR Dataset/train/X_train.txt")
   
6. use.test.x <- read.table("UCI HAR Dataset/test/X_test.txt")
   
7. 
   
8. use.train.y <- read.table("UCI HAR Dataset/train/y_train.txt")[[1]]
   
9. use.test.y <- read.table("UCI HAR Dataset/test/y_test.txt")[[1]]
   
10. 
    
11. use.train <- cbind(use.train.x, Outcome = factor(use.train.y))
    
12. use.test <- cbind(use.test.x, Outcome = factor(use.test.y))
    
13. 
    
14. use.labels <- read.table("UCI HAR Dataset/activity_labels.txt")
    
15. 
    
16. h2oactivity.train <- as.h2o(
    
17.   use.train,
    
18.   destination_frame = "h2oactivitytrain")
    
19. 
    
20. h2oactivity.test <- as.h2o(
    
21.   use.test,
    
22.   destination_frame = "h2oactivitytest")
    
23. We have already learned the components of training a deep prediction model. We use the h2o.deeplearning() function as we did for the auto-encoder models, but specify the variable names for both the x and y arguments. Before, we included the testing data to automatically get performance metrics on both training and testing data. However, to show how to generate predictions on new data, we do not include it in the call to h2o.deeplearning(). The activation function used is a linear rectifier with dropout both on the input variables (20%) and the hidden neurons (50%). This little example is a shallow network with only 50 hidden neurons and 10 training iterations. The cost (loss) function is cross-entropy:
    
24. 
    
25. mt1 <- h2o.deeplearning(
    
26.   x = colnames(use.train.x),
    
27.   y = "Outcome",
    
28.   training_frame= h2oactivity.train,
    
29.   activation = "RectifierWithDropout",
    
30.   hidden = c(50),
    
31.   epochs = 10,
    
32.   loss = "CrossEntropy",
    
33.   input_dropout_ratio = .2,
    
34.   hidden_dropout_ratios = c(.5), ,
    
35.   export_weights_and_biases = TRUE
    
36. )

复制代码