Learning Hidden Markov Models for Regression using Path Aggregation
This addresses regression tasks for sequential data, but appears incremental as it combines existing HMM and regression techniques.
The paper tackles the problem of learning mappings from sequential data to real-valued responses by jointly learning a hidden Markov model (HMM) and a regression model, demonstrating its value in synthetic and biological domains.
We consider the task of learning mappings from sequential data to real-valued responses. We present and evaluate an approach to learning a type of hidden Markov model (HMM) for regression. The learning process involves inferring the structure and parameters of a conventional HMM, while simultaneously learning a regression model that maps features that characterize paths through the model to continuous responses. Our results, in both synthetic and biological domains, demonstrate the value of jointly learning the two components of our approach.