0

I want to implement a real-time recommendation on Apache Flink with the ALS algorithm.

The model can be previously trained ready with Batch and then just loaded into Flink. Then an input stream of Data should be processed and used for the prediction.

The library FlinkMl is no longer included since Flink 1.9. Beyond that, there are a lot of libraries designed for machine learning with Apache Flink.

For this I need a starting point that helps me to realize this project.

Olaf Kock
  • 46,930
  • 8
  • 59
  • 90
jkempter
  • 3
  • 2

1 Answers1

0

1、Recommend to try Alink.Alink is the Machine Learning algorithm platform based on Flink, developed by the PAI team of Alibaba computing platform.it supports the ALS algorithm. https://github.com/alibaba/Alink/blob/master/README.en-US.md

you’ll find the README that can explain it much better.

2、Try Pandas?Pandas UDF/UDAF can be called directly in PyFlink.

hope it can help u.

liliwei
  • 294
  • 1
  • 8
  • Do you know an example Repository that is using "alink" as ML library? – jkempter Feb 20 '21 at 09:56
  • you can find example and tutorial on github .(https://github.com/alibaba/Alink/tree/master/examples/src/main/java/com/alibaba/alink) – liliwei Feb 20 '21 at 10:34