Machine Learning algorithms for data analysis makes new areas such as Big Data analysis. Emerging data analysis with the streaming data is one of the most paramount areas that will be pioneered the businesses in the future. WSO2 ML (Machine Learner) is one of the main platforms which building some intelligent and predictive data analysis based on the machine learning algorithms to support modern data transactions. Even though they have ML which is build upon the well known Apache Spark MLLib, WSO2 ML still could not support streaming data. Even though the Spark MLLib has streaming support for k mean clustering and generalized linear regression (GLR) , it developed with a Scala API. Purpose of this project is to develop a Java API to support streaming k mean clustering and GLR with the mini batch sampling techniques to support streaming data without using Spark streaming. Therefore design include a way to acquire streaming data and break those data streams into mini batches which can be used this batches to retrain the models periodically with some optimization techniques such as Stochastic gradient descent algorithms.



mahesh dananjaya


  • Supun
  • maheshakya
  • Nirmal Fernando