Gauss Labs pre test

Introduction

• Review the paper "Efficient Learning Algorithms for Changing Environment : https://icml.cc/Conferences/2009/papers/75.pdf
• Implement the "Follow-the-Leading-History(FLH)" algorithms
• use dataset which incluldes sensor value and measurement value obtained from manufacturing process and product

Detail

• Programming Language / Package : Python, Numpy, Scikit-learn(metrics)
• Algorithms for implementation
  1. Modified Ridge Regression
  2. Aggregating Algorithm for Regression
  3. Follow-the-Leading-History(FLH) algorithm

Result

Prediction values for every time step of data with AAR and FLH algorithms

Reference

• Efficient Learning Algorithms for Changing Environments : https://icml.cc/Conferences/2009/papers/75.pdf
• Adaptive Algorithms for Online Decision Problems : https://www.researchgate.net/publication/220138797_Adaptive_Algorithms_for_Online_Decision_Problems
• Non-stationary Online Regression : https://arxiv.org/abs/2011.06957
• Online Ridge Regression Method Using Sliding Windows : https://www.researchgate.net/publication/236648230_Online_Ridge_Regression_Method_Using_Sliding_Windows