Stacked Generalization with Wrapper-based Feature Selection for Human Activity Recognition

Introduction

1.Human Activity Recognition

2.Ensemble Learning

3.Feature Selection using Boruta

Applications

1.Ambient Living and Assistance

2.Elderly care and support

3.Monitoring and surveillance

Previous Research

Experimental methods and observations

1. Dataset Pre-processing

2. Feature Selection using Boruta

3. Model: Stacked ensemble

4. Training and Observations

Conclusions

1.Improved performance

2.Viability of ensemble models for various situations

3.Future work