How do I use ML models to estimate current stress level based on past data?

Question

I am new to machine learning and I cannot understand the difference between estimating current stress level and predicting future stress levels based on historical data. I have been told these are two different problems and require a different approach. I have a dataset with the features and the stress level column, which is the target. Now if I want to estimate the current stress level does this mean I have to generate lag-based features? and then use a time-split for training and testing?

Answer 1

Yes, both these scenarios are different.

1. Estimating Current Stress level - Your target variable here is stress level and features are heart rate and blood pressure. In order to estimate current stress level; you would have the values of the features and they would be known to you as it is in present and you can calculate heart rate and blood pressure using devices. So, you can directly apply machine learning models or even use operational research mathematical models to calculate the stress level.

2. Estimating future stress level - Your target variable here is stress level and features here are still heart rate and blood pressure. In order to estimate future stress level; you wouldn't know what would be the future heart rates and blood pressure; those will be unknown to you. The extra step here would be to impute these unknown values. So, you would have to consider different imputation techniques for these features for the future; some common concepts that you can utilize are imputing the lag values for both heart rate and blood pressure and then utilizing the dataset for your machine learning models.