Machine Learning Projects for .NET Developers Apress 2015

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

Machine Learning Projects for .NET Developers

Book Description
Machine Learning Projects for .NET Developers shows you how to build smarter .NET applications that learn from data, using simple algorithms and techniques that can be applied to a wide range of real-world problems. You'll code each project in the familiar setting of Visual Studio, while the machine learning logic uses F#, a language ideally suited to machine learning applications in .NET. If you're new to F#, this book will give you everything you need to get started. If you're already familiar with F#, this is your chance to put the language into action in an exciting new context.

Along the way, you'll learn fundamental ideas that can be applied in all kinds of real-world contexts and industries, from advertising to finance, medicine, and scientific research. While some machine learning algorithms use fairly advanced mathematics, this book focuses on simple but effective approaches. If you enjoy hacking code and data, this book is for you.
Book Details
Publisher:        Apress
By:        Mathias Brandewinder
ISBN:        978-1-4302-6767-6
Year:        2015
Pages:        300
Language:        English
File size:        7.5 MB
File format:        PDF
eBook
Download:        Machine Learning Projects for .NET Developers

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5 回帖

关键词：Developers developer Learning Projects earning techniques learning familiar problems machine

1. Listing 2-1. Reading the SMS dataset from the file
   
2. open System.IO
   
3. type DocType =
   
4. | Ham
   
5. | Spam
   
6. let parseDocType (label:string) =
   
7. match label with
   
8. | "ham" -> Ham
   
9. | "spam" -> Spam
   
10. | _ -> failwith "Unknown label"
    
11. let parseLine (line:string) =
    
12. let split = line.Split('\t')
    
13. let label = split.[0] |> parseDocType
    
14. let message = split.[1]
    
15. (label, message)
    
16. Let fileName = "SMSSpamCollection"
    
17. let path = __SOURCE_DIRECTORY__ + @"..\..\Data\" + fileName
    
18. let dataset =
    
19. File.ReadAllLines path
    
20. |> Array.map parseLine
    
21. At this point, you should be able to select the entire code you just wrote in the script file and execute it
    
22. in F# Interactive, which will produce the following:
    
23. val dataset : (DocType * string) [] =
    
24. [|(Ham,
    
25. "Go until jurong point, crazy.. Available only in bugis n grea"+[50 chars]);
    
26. (Ham, "Ok lar... Joking wif u oni...");
    
27. (Spam,
    
28. "Free entry in 2 a wkly comp to win FA Cup final tkts 21st May"+[94 chars]);
    
29. // Snipped for brevity
    
30. (Ham,
    
31. "Hi. Wk been ok - on hols now! Yes on for a bit of a run. Forg"+[117 chars]);
    
32. (Ham, "I see a cup of coffee animation"); ...|]

复制代码

本帖隐藏的内容

Machine Learning Projects for .NET Developers.rar (5.51 MB, 需要: 20 个论坛币) 本附件包括：

• Machine Learning Projects for .NET Developers.pdf

2015-8-2 23:11:10 上传

需要: 20 个论坛币  [购买]

Assuming a document has already been broken into a set of tokens, computing its score for a given group is fairly simple. We don’t have to worry about how these numbers have been computed. If that part has been done already, all we need to do is sum up the log of the group frequency, as well as the log of every token that is present in both the tokenized document and the list of tokens we are using in our model:

1. Listing 2-2. Scoring a document
   
2. let tokenScore (group:DocsGroup) (token:Token) =
   
3. if group.TokenFrequencies.ContainsKey token
   
4. then log group.TokenFrequencies.[token]
   
5. else 0.0
   
6. let score (document:TokenizedDoc) (group:DocsGroup) =
   
7. let scoreToken = tokenScore group
   
8. log group.Proportion +
   
9. (document |> Seq.sumBy scoreToken)

复制代码

1. Listing 2-3. Predicting the label of a document
   
2. let classify (groups:(_*DocsGroup)[])
   
3. (tokenizer:Tokenizer)
   
4. (txt:string) =
   
5. let tokenized = tokenizer txt
   
6. groups
   
7. |> Array.maxBy (fun (label,group) ->
   
8. score tokenized group)
   
9. |> fst

复制代码

1. Listing 2-4. Analyzing a group of documents
   
2. let analyze (group:TokenizedDoc seq)
   
3. (totalDocs:int)
   
4. (classificationTokens:Token Set)=
   
5. let groupSize = group |> Seq.length
   
6. let score token =
   
7. let count = countIn group token
   
8. laplace count groupSize
   
9. let scoredTokens =
   
10. classificationTokens
    
11. |> Set.map (fun token -> token, score token)
    
12. |> Map.ofSeq
    
13. let groupProportion = proportion groupSize totalDocs
    
14. {
    
15. Proportion = groupProportion
    
16. TokenFrequencies = scoredTokens
    
17. }

复制代码

1. Listing 2-5. Learning from documents
   
2. let learn (docs:(_ * string)[])
   
3. (tokenizer:Tokenizer)
   
4. (classificationTokens:Token Set) =
   
5. let total = docs.Length
   
6. docs
   
7. |> Array.map (fun (label,docs) -> label,tokenizer docs)
   
8. |> Seq.groupBy fst
   
9. |> Seq.map (fun (label,group) -> label,group |> Seq.map snd)
   
10. |> Seq.map (fun (label,group) -> label,analyze group total classificationTokens)
    
11. |> Seq.toArray

复制代码

1. Listing 2-6. Training a naïve Bayes classifier
   
2. let train (docs:(_ * string)[])
   
3. (tokenizer:Tokenizer)
   
4. (classificationTokens:Token Set) =
   
5. let groups = learn docs tokenizer classificationTokens
   
6. let classifier = classify groups tokenizer
   
7. classifier

复制代码

1. Listing 2-7. Using regular expressions to tokenize a line of text
   
2. open System.Text.RegularExpressions
   
3. let matchWords = Regex(@"\w+")
   
4. let tokens (text:string) =
   
5. text.ToLowerInvariant()
   
6. |> matchWords.Matches
   
7. |> Seq.cast<Match>
   
8. |> Seq.map (fun m -> m.Value)
   
9. |> Set.ofSeq

复制代码

1. Listing 2-8. A generic model-evaluation function
   
2. let evaluate (tokenizer:Tokenizer) (tokens:Token Set) =
   
3. let classifier = train training tokenizer tokens
   
4. validation
   
5. |> Seq.averageBy (fun (docType,sms) ->
   
6. if docType = classifier sms then 1.0 else 0.0)
   
7. |> printfn "Correctly classified: %.3f"
   
8. Evaluating a particular model then becomes a one-liner:
   
9. > evaluate wordTokenizer allTokens;;

复制代码