Time series machine learning techniques show great promise for the analysis of health care wearable data. As our busy lifestyles render continuous monitoring more and more essential, the need to analyze data to find correlations between these data streams becomes even more important, because they can provide important cues to people. These cues could be as simple as reminding a person to take a walk or move around, which is already being done by a lot of wearables available today, such as Fitbit, Garmin, Nike, etc. However, along with monitoring the current state of an individual, these popular devices are not able to perform the complex predictions that correlate the captured information to make sense at a higher level or provide causal relationships between the data. My research aims to develop advanced algorithms for analyzing time series data for estimation and prediction of physiological parameters (such as heart rate or respiration rate using kinematic and physiological data).