F#

and

Azure Machine Learning

jj09.net

about me

              swim

 

 

 

 bike

              run

F#

Mac

fsharpi

fsharpc

VSCode(Ionide)

PC

fsi

fsc

VS(Code)

Machine Learning

Learning Techniques

  • Supervised learning
     

  • Unsupervised learning

Supervised learning

  • Classification problems
     

  • Regression problems

Machine Learning is hard

  • whorepresents.com

  • therapistfinder.com

  • expertsexchange.com

  • penisland.com

  • speedofart.net

Stock Estimator

Linear Regression

Code

github.com/jj09/StockEstimator

Azure

Machine Learning

Stock Estimator

 

 

What next?

 

Deploy to Azure

  • API: Docker + Fake = <3

    • stockestimator.westus.cloudapp.azure.com
  • UI: App Service

    • stockestimator.azurewebsites.net
  • using Azure Service Fabric?

  • bonus: gcloud
    • stockestimator.appspot.com

Add more algorithms

  • Boosted Decision Tree Regression?

  • Neural Network Regression?

  • Decision Forest Regression?

misc

  • beautiful UI

    • chart.js?
    • plotly.js?
    • smoothie.js?
    • three.js?
  • real-time estimation

  • perf optimizations

    • 1 request per estimated range

    • cache data on server and client

    • upgrade to ASP.NET Core 1.1 and Aurelia 1.0

Learn more

Websites

Pluralsight

People

Machine Learning

Azure Machine Learning

Stay in touch!

 

jj09.net

 

@JakubJedryszek

F# and Azure Machine Learning

By Jakub Jedryszek

F# and Azure Machine Learning

F# is not a replacement, but a great complement for C#. Currently F# is used in many financial applications. Let’s see how we can predict future stock prices with power of F#, C#, and Azure Machine Learning. In this session I will show you how to build F# Backend for estimating future stock prices with Azure Machine Learning, how to create Web API powered by F# with Suave Framework, and how to consume it from your ASP.NET Core App with Front-End powered by Aurelia and D3.js.

  • 2,911