Infragistics WPF controls

numl - a machine learning library for .NET developers

In one of my previous posts called Machine learning resources for .NET developers, I introduced a machine learning library called numl.net.  numl.net is a machine learning library for .NET created by Seth Juarez.  You can find the library here and Seth's blog here.  When I began researching the library, I learned quickly that one of Seth's goals in writing numl.net was to abstract away the complexities that stops many software developers from trying their hand at machine learning.  I must say that in my opinion, he has done a wonderful job in accomplishing this goal! Tutorial I've decided to throw together a small tutorial to show you just how easy it is to use numl.net to perform predictions.  This tutorial will use structured learning by way of a decision tree to perform predictions.  I will use the infamous Iris Data set which contains data 3 different types of Iris flowers and the data that defines them.  Before we get into code, let's look at some basic terminology first. With numl.net you create a POCO (plain old CLR object) to use for training as well as predictions.  There will be properties that you will specify known values (features) so that you can predict the value of an unknown property value (label).  numl.net makes identifying features and labels easy, you simply mark your properties with the [Feature] attribute or the [Label] attribute (there is also a [StringLabel] attribute as well).  Here is an example of the Iris class that we will use in this tutorial. using numl.Model; using System; using System.Collections.Generic; using System.Linq; using System.Text; using System.Threading.Tasks; namespace NumlDemo { /// <summary> /// Represents an Iris in the infamous Iris classification dataset (Fisher, 1936) /// Each feature property will be used for training as well as prediction. The label /// property is the value to be predicted. In this case, it's which type of Iris we are dealing with. /// </summary> public class Iris { //Length in centimeters [Feature] public double SepalLength { get; set; } //Width in centimeters [Feature] public double SepalWidth { get; set; } //Length in centimeters [Feature] public double PetalLength { get; set; } //Width in centimeters [Feature] public double PetalWidth { get; set; } //-- Iris Setosa //-- Iris Versicolour //-- Iris Virginica public enum IrisTypes { IrisSetosa, IrisVersicolour, IrisVirginica } [Label] public IrisTypes IrisClass { get; set; } //This is the label or value that we wish to predict based on the supplied features } } As you can see, we have a simple POCO Iris class, which defines four features and one label.  The Iris training data can be found here .  Here is an example of the data found in the file.   5.1,3.5,1.4,0.2,Iris-setosa 6.3,2.5,4.9,1.5,Iris-versicolor 6.0,3.0,4.8,1.8,Iris-virginica     The first four values are doubles which represent the features Sepal Length, Sepal Width, Petal Length, Petal Width.  The final value is an enum that represents the label that we will predict which is the class of Iris.   We have the Iris class, so now we need a method to parse the training data file and generate a static List<Iris> collection.  Here is the code:   using System; using System.Collections.Generic; using System.IO; using System.Linq; using System.Text; using System.Threading.Tasks; namespace NumlDemo { /// <summary> /// Provides the services to parse the training data files /// </summary> public static class IrisDataParserService { //provides the training data to create the predictive model public static List<Iris> TrainingIrisData { get; set; } /// <summary> /// Reads the trainingDataFile and populates the TrainingIrisData list /// </summary> /// <param name="trainingDataFile">File full of Iris data</param> /// <returns></returns> public static void LoadIrisTrainingData(string trainingDataFile) { //if we don't have a training data file if (string.IsNullOrEmpty(trainingDataFile)) throw new ArgumentNullException("trainingDataFile"); //if the file doesn't exist on the file system if (!File.Exists(trainingDataFile)) throw new FileNotFoundException(); if (TrainingIrisData == null) //initialize the return training data set TrainingIrisData = new List<Iris>(); //read the entire file contents into a string using (var fileReader = new StreamReader(new FileStream(trainingDataFile, FileMode.Open))) { string fileLineContents; while ((fileLineContents = fileReader.ReadLine()) != null) { //split the current line into an array of values var irisValues = fileLineContents.Split(','); double sepalLength = 0.0; double sepalWidth = 0.0; double petalLength = 0.0; double petalWidth = 0.0; if (irisValues.Length == 5) { Iris currentIris = new Iris(); double.TryParse(irisValues[0], out sepalLength); currentIris.SepalLength = sepalLength; double.TryParse(irisValues[1], out sepalWidth); currentIris.SepalWidth = sepalWidth; double.TryParse(irisValues[2], out petalLength); currentIris.PetalLength = petalLength; double.TryParse(irisValues[3], out petalWidth); currentIris.PetalWidth = petalWidth; if (irisValues[4] == "Iris-setosa") currentIris.IrisClass = Iris.IrisTypes.IrisSetosa; else if (irisValues[4] == "Iris-versicolor") currentIris.IrisClass = Iris.IrisTypes.IrisVersicolour; else currentIris.IrisClass = Iris.IrisTypes.IrisVirginica; IrisDataParserService.TrainingIrisData.Add(currentIris); } } } } } } This code is pretty standard.  We simply read each line in the file, split the values out into an array, and populate a List<Iris> collection of Iris objects based on the data found in the file.   Now the magic Using the numl.net library, we need only use three classes to perform a prediction based on the Iris data set.  We start with a Descriptor, which identifies the class in which we will learn and predict.  Next, we will instantiate a DecisionTreeGenerator, passing the descriptor to the constructor.  Finally, we will create our prediction model by calling the Generate method of the DecisionTreeGenerator, passing the training data (IEnumerable<Iris>) to the Generate method.  The generate method will provide us with a model in which we can perform our prediction. Here is the code: using numl; using numl.Model; using numl.Supervised; using System; using System.Collections.Generic; using System.Linq; using System.Text; using System.Threading.Tasks; namespace NumlDemo { class Program { public static void Main(string[] args) { //get the descriptor that describes the features and label from the Iris training objects var irisDescriptor = Descriptor.Create<Iris>(); //create a decision tree generator and teach it about the Iris descriptor var decisionTreeGenerator = new DecisionTreeGenerator(irisDescriptor); //load the training data IrisDataParserService.LoadIrisTrainingData(@"D:\Development\machinelearning\Iris Dataset\bezdekIris.data"); //create a model based on our training data using the decision tree generator var decisionTreeModel = decisionTreeGenerator.Generate(IrisDataParserService.TrainingIrisData); //create an iris that should be an Iris Setosa var irisSetosa = new Iris { SepalLength = 5.1, SepalWidth = 3.5, PetalLength = 1.4, PetalWidth = 0.2 }; //create an iris that should be an Iris Versicolor var irisVersiColor = new Iris { SepalLength = 6.1, SepalWidth = 2.8, PetalLength = 4.0, PetalWidth = 1.3 }; //create an iris that should be an Iris Virginica var irisVirginica = new Iris { SepalLength = 7.7, SepalWidth = 2.8, PetalLength = 6.7, PetalWidth = 2.0 }; var irisSetosaClass = decisionTreeModel.Predict<Iris>(irisSetosa); var irisVersiColorClass = decisionTreeModel.Predict<Iris>(irisVersiColor); var irisVirginicaClass = decisionTreeModel.Predict<Iris>(irisVirginica); Console.WriteLine("The Iris Setosa was predicted as {0}", irisSetosaClass.IrisClass.ToString()); Console.WriteLine("The Iris Versicolor was predicted as {0}", irisVersiColorClass.IrisClass.ToString()); Console.WriteLine("The Iris Virginica was predicted as {0}", irisVirginicaClass.IrisClass.ToString()); Console.ReadKey(); } } } And that's all there is to it.  As you can see, you can use the prediction model accurately and there's no math, only simple abstractions. I hope this has peaked your interest in the numl.net library for machine learning in .NET.   Feel free to post any questions or opinions. Thanks for reading! Buddy James  


Machine learning resources for .NET developers

Greetings friends and welcome to this article on Machine learning libraries for .NET developers.  Machine learning is a hot topic right now and for good reason.  Personally, I haven't been so excited about a technology since my computer used my 2800 baud modem to dial into a BBS over 17 years ago.  The thought that my computer could communicate with another computer was so fascinating to me.  That moment was the very moment that would forever change my life.  I learned a lot about DOS by writing batch scripts and running other programs that allowed me to visit and then run a BBS system.  It eventually lead me to QBasic.  I wanted to learn to write BBS door games and QBasic was included as a part of a standard DOS installation back then. Fast forward 17 years and I'm still in love with computers, programming, and the concept of communication between machines.  The magic never disappeared.  So when i first learned about the concept of Machine learning, I felt like that 13 year old kid again.  The idea that a machine can learn to do things that it has not been programmed to do is now a passion of mine.  The concepts of Machine learning have an extreme learning curve, however, I believe that we as humans can do anything that we put our mind to.  So I began looking around for tutorials on machine learning.  I found many great tutorials and books, however, most of them involved using python.  I have nothing against python.  As a matter of fact, I find it ironic that I started with BASIC and now in this moment of "rebirth" I'm beginning to use python which looks a lot like BASIC in many ways.  The fact of the matter remains, I'm a .NET developer.  I've spent the last 9 years in the .NET framework and I love the technology.  C# is an awesome programming language and it's hard to imagine life without Visual Studio.  What can I say, the IDE has spoiled me. While I scoured the internet looking for tutorials related to Machine learning resources for .NET developers, I wished that there was a one resource that would assist me in my search for resources to help me achieve my goal. Well that's what this article is all about.  In this article, I will introduce you to some .NET libraries that will assist you in your quest to learn about Machine learning. NND Neural Network Designer by Bragisoft The Neural Network Designer project (NND) is a DBMS management system for neural networks that was created by Jan Bogaerts.  The designer application is developed using WPF, and is a user interface which allows you to design your neural network, query the network, create and configure chat bots that are capable of asking questions and learning from your feed back.  The chat bots can even scrape the internet for information to return in their output as well as to use for learning.  The project includes a custom language syntax called NNL (neural network language) that you can use in configuring your machine learning project.  The source code is designed so that the libraries can be used in your own custom applications so you don't have to start from scratch with such a complex set of technologies.  The project is actually an open source project in which I am a part of.  Some of the possibilities offered by this awesome project include predictions, image and pattern recognition, value inspection, memory profiling and much more.  Stop by the Bragisoft NND website and download the application to give it a try.   Screen shots of the neural network designer by Bragisoft A DBMS for neural networks   Mind map rand forrest The chat bot designer and other tools Accord.net Here is a description from the Accord.NET project website  Accord.NET is a framework for scientific computing in .NET. The framework builds upon AForge.NET, an also popular framework for image processing, supplying new tools and libraries. Those libraries encompass a wide range of scientific computing applications, such as statistical data processing, machine learning, pattern recognition, including but not limited to, computer vision and computer audition. The framework offers a large number of probability distributions, hypothesis tests, kernel functions and support for most popular performance measurements techniques.  The most impressive parts of this library has got to be the documentation and sample applications that are distributed with the project.  This makes the library easy to get started using.  I also like the ability to perform operations like Audio processing (beat detection and more), Video processing (easy integration with your web cam, vision capabilities and object recognition).  This is an excellent place to start with approaching Machine learning with the .NET framework.  Here are a two videos that should whet your appetite. Hand writing recognition with Accord.NET   Here is an example of head tracking with Accord.NET (super cool)   AIMLBot Program# AILM Chat bot library AIMLBot (Program#) is a small, fast, standards-compliant yet easily customizable implementation of an AIML (Artificial Intelligence Markup Language) based chatter bot in C#. AIMLBot has been tested on both Microsoft's runtime environment and Mono. Put simply, it will allow you to chat (by entering text) with your computer using natural language.  The project is located here. Math.NET Machine learning algorithms are extremely math heavy.  Math.NET is a library  that can assist with the math that is required to solve machine learning related problems. Math.NET Numerics aims to provide methods and algorithms for numerical computations in science, engineering and every day use. Covered topics include special functions, linear algebra, probability models, random numbers, interpolation, integral transforms and more. DotNumerics DotNumerics is a website dedicated to numerical computing for .NET. DotNumerics includes a Numerical Library for .NET. The library is written in pure C# and has more than 100,000 lines of code with the most advanced algorithms for Linear Algebra, Differential Equations and Optimization problems. The Linear Algebra library includes CSLapack, CSBlas and CSEispack, these libraries are the translation from Fortran to C# of LAPACK, BLAS and EISPACK, respectively. You can find the library here.  ALGLIB ALGLIB is a cross-platform numerical analysis and data processing library. It supports several programming languages (C++, C#, Pascal, VBA) and several operating systems (Windows, Linux, Solaris). ALGLIB features include: Accessing ‘R’ from C#–Lessons learned Here are instructions to use the R statistical framework from within c# ILNumerics You can check out the library at http://www.ilnumerics.net NuML.net http://numl.net A nice site about the basics of machine learning in c# by Seth Juarez . NuML.NET is a machine learning library for .NET developers written by Seth Juarez.  I've recently tried this library and I'm impressed!  Seth has stated publicly that his intention behind the numl.net library is to abstract the scary math away from machine learning to provide tools that are more approachable by software developers and boy did he deliver!  I've been working with this library for a little more than an hour and I've written a prediction app in c#.  You can find his numl.net library source on github. Encog Machine Learning Framework Here is what the official Heaton Research website has to say about Encog: Encog is an advanced machine learning framework that supports a variety of advanced algorithms, as well as support classes to normalize and process data. Machine learning algorithms such as Support Vector Machines, Artificial Neural Networks, Genetic Programming, Bayesian Networks, Hidden Markov Models and Genetic Algorithms are supported. Most Encog training algoritms are multi-threaded and scale well to multicore hardware. Encog can also make use of a GPU to further speed processing time. A GUI based workbench is also provided to help model and train machine learning algorithms. Encog has been in active development since 2008. Encog is available for Java, .Net and C/C++. Jeff Heaton knows a great deal about machine learning algorithms and he's created a wonderful library called Encog.  I was able to write a neural network application that solved the classic XOR problem in 20 minutes after installing the library.  What really amazes me is that he has an Encog Library for JavaScript which includes live samples on his website of Javascript + encog solving problems like the Traveling Salesman Problem and Conway's game of life, all in a browser!  This library can even use your GPU for the heavy lifting if that's your choice.  I would highly recommend that you at least check out his site and download the library to look at the examples.  You can find the Encog library here.    Conclusion This concludes my article on Machine learning resources for the .NET developer.  If you have any suggestions regarding a project that you know of or you are working on related to Machine learning in .NET, please don't hesitate to leave a comment and I will update the article to mention the project.  This article has shown that we as .NET developers have many resources available to us to use to implement Machine learning based solutions.  I appreciate your time in reading this article and I hope you found it useful.  Please subscribe to my RSS feed.  Until next time.. Buddy James


Machine Learning: 5 examples of what it is and why you should care

Machine learning examples to make you think Hello folks, and welcome to another awe inspiring article from refactorthis.net .  This article is one that I'm very excited to present.  I'm sure you've guessed by now that the topic of this post is Machine learning.   If you don't know what machine learning is or don't care, I ask you to take a look at the fascinating examples that I've presented in this article.  You just may get inspired.   Background I'm a .NET developer and I have experience working in a myriad of different business domains.  My love affair with machine learning was brought about while I was working for an e-commerce website.  My boss had asked me to look into a new API that Google had released in beta at the time that would allow you to provide data about your customers and it would suggest products based on their shopping data.  That API is called Google prediction and you can read about the Google prediction API here.  It's been a couple of years since I was introduced to this technology and since that time I've contemplated how machine learning algorithms work and what possibilities they could unlock with the right amount of data and creativity. Since then I've done a lot of reading, and planning on ways that I can collect data to use in my journey to learn as much as I can about this new frontier that we as software developers are facing.  I was fortunate enough become the first accepted team member of the open source project called NND, or Neural Network Designer by Bragisoft.  Check back soon for an article dedicated to this wonderful open source project on machine learning and neural networks. Proceed with caution Let it be known, however, that machine learning algorithms are not for the faint of heart.  This is a very complex array of concepts and I don't plan to  try to explain them in this article.  What I will do is give a brief, simple introduction to a few of the prevalent topics that one would need to research in order to implement machine learning algorithms.  The main purpose of this article is to provide some wonderful youtube videos that provide insight into the possibilities of machine learning and it's practical applications.  Some of the videos just may blow your mind! So without further ado, let's bring forth the videos! A glimpse at the future to whet your appetite  This first video is full of commentary and stunning examples of robots that lack brains but are capable of learning by way of a design that mimics a central nervous system.  Although the video depicts the future as something to be worried about, it's still a great, non technical introduction to whet your appetite and make way for the other more specific videos that are focused on simulations and applications of machine learning. Watch a simulation of robots that learn to drive In this video, we see a software demonstration of animated robots that start with no knowledge of a driving course.  Watch what happens as the simulation is processed repeatedly, allowing the robots to learn the course with each generation.  Buckle up! Hey Darwin, what do you think of this? This next video is progressively more complex, however, the simulation clearly shows how a collection of bots act as they evolve.  The bots  can attack each other with the intent of killing and eating each other to sustain life.  They also reproduce to make baby bots.  Take a look at this awesome example of machine learning.. Virtual Darwinism take two This next video is another representation of virtual evolution by machine learning techniques.  Watch block like organisms learn to fight each other over a virtual cube that represents a block of food.   Machine learning for pattern matching and recognition This video shows an application that will render CAPTCHA verification methods obsolete.  I have a decent amount of experience working with OCR engines however, this is the most accurate recognition I've ever seen. Show and tell This concludes my article on machine learning examples.  I hope you enjoyed the article and that it has peaked your interest in machine learning.  Do you have a favorite video or application that demonstrates machine learning, neural networks, prediction algorithms, pattern matching or some other related technology?  If so, we'd love to hear about it so please leave a comment with a link. Thanks for reading, Buddy James kick it on DotNetKicks.com


Complete coverage of your source code with NDepend part 1

What is NDepend? This article is part one of a two part series about one of the most practical and dynamic tools in existence for .NET development.  I’m talking about NDepend http://www.NDepend.com.  I was approached about writing a review for NDepend so I downloaded the application to give it a try.  As with all of my reviews, let it be known that if I think a product is mediocre, then that’s what I’m going to write.  All that to say that this is no exaggeration, I really feel this strongly about this tool.  I’m sure by the end of this article, I will have peeked your interest too.  If you are interested, please read on. NDepend pro product suite From NDepend.com, “NDepend is a Visual Studio tool to manage complex .NET code and achieve high Code Quality.”  This tool allows you to visualize your source code in many different ways in an effort to analyze the quality of your code and how to improve it.  The product comes complete with a Visual Studio add in, an independent GUI tool, and a set of power tools that are console based which makes the product suite extremely versatile.  Whether you are pressed for time and need to analyze your code while in visual studio, you prefer a standalone GUI, or you are addicted to the command line, this product is made to fit your needs. Installation The NDpend installation process is very straight forward.  The download is a zip file that contains the complete product suite.  You simply pick a folder to install to and unzip the archive.  If you’ve purchased the pro version, you will be provided with a license in the form of an XML file which needs to be placed in the directory that you chose to install the product. Installing the Visual Studio 2012 add-in Once you’ve unzipped the archive, you need to run the NDepend.Install.VisualStudioAddin.exe executable to install the Visual Studio add-in. Running the install The installation completed Adding an NDepend project to your solution When you use the Visual Studio integration, you need to create an NDepend project in the solution that you wish to analyze. NDepend will tell you anything that wish you know about source code.  This is powerful, however, it’s a point that must be covered.  In order to be productive with NDepend, you must first define what information that you wish to discover about your source code and how you plan to use that information.  If you don’t have this information then you will not get much use from the product.  The information that it provides to you is very useful, however, you must take some time to plan out how you will use this information to benefit you and your coding efforts. You may wish to make sure that your code maintains a consistent amount of test coverage.  Perhaps you wish to make sure that all methods in your codebase stay below a certain threshold regarding the number of lines of code that they contain.  NDepend is capable of telling you this and much more about your source code. One of the coolest features that I’ve seen in the product is the Code Query Linq (CQLinqing).  This allows you to query your source code using LINQ syntax to bring back anything that you wish to know about your source code.   You can query an assembly, a class, even a method.  The product comes with predefined CQLinq rules but also allows you to create your own rules as well as edit existing rules. I plan to write another blog post that explains my personal experience with the product.  I’ve recently joined an open source project that is a framework that handles some very advanced topics such Artificial intelligence, Machine learning, and language design.  The project is called neural network designer http://bragisoft.com/ .  I chose this project because the source code is vast and I believe that a large code base is a perfect target to use NDepend to get the most benefit. I plan to use the product and test the following areas:   What information do I want to know about my code base?   When do I wish to be presented with this information?   How do I plan on using this information to improve my code?   How can I use NDepend to provide this information? I think that if you wish to get any use out of the product, it will be very important that you answer these questions.  The product is vast and diverse but it can also be a bit intimidating.  With that said, I plan to use my next post to illustrate how I was able to use NDepend to define the metrics that I needed from my code, and how I used NDepend to provide those metrics to me. Stay tuned for the next installment which will explain my experience with using NDepend to improve my development efforts and my source code. Thanks for reading, Buddy James kick it on DotNetKicks.com


About the author

My name is Buddy James.  I'm a Microsoft Certified Solutions Developer from the Nashville, TN area.  I'm a Software Engineer, an author, a blogger (http://www.refactorthis.net), a mentor, a thought leader, a technologist, a data scientist, and a husband.  I enjoy working with design patterns, data mining, c#, WPF, Silverlight, WinRT, XAML, ASP.NET, python, CouchDB, RavenDB, Hadoop, Android(MonoDroid), iOS (MonoTouch), and Machine Learning. I love technology and I love to develop software, collect data, analyze the data, and learn from the data.  When I'm not coding,  I'm determined to make a difference in the world by using data and machine learning techniques. (follow me at @budbjames).  

Related links

Month List