ML.NET (Machine Learning for .NET) is an Open Source and Cross Platform Framework (Windows, Linux, macOS) which makes machine learning accessible for .NET developers.

ML.NET allows us to build machine learning models to get scenarios like sentiment analysis, fraud detection, forecasting, issue classification, image classification, recommendations & many more.

We have a new announcement for ML.NET 0.8 on December 4, 2018, and ML.NET 0.8 is an improved version of ML.NET 0.7. One great feature of ML.NET 0.8 is One Class Matrix Factorization which allows us to build recommendations like Frequently bought together. A code sample is on GitHub One Class Matrix Factorization.

A Matrix Factorization is a supervised machine learning. Steps to solve the problems is shown below in the diagram,

Some more new features are,

• Improved debuggability by previewing the data
• Model explainability

API improvements in ML.NET 0.8

• Filtering rows in a DataView

  • IDataView trainingDataView = mlContext.Data.FilterByColumn(baseTrainingDataView, "FareAmount", lowerBound: 1, upperBound: 150);

     

• Caching APIs
• Enabled saving and loading data in IDataView (IDV) binary format for improved performance
• Enabled stateful prediction engine for time series problems such as anomaly detection

Get started

If you haven’t already get started with ML.NET here.