How to create a single value using time series data set?

Question

I have following data sets(265 data sets). Its a heart beat data set which is collected in every 1 second interval.

**Heart rate pattern**
82 82 87 87 89 90 89 89 89 89
88 89 89 87 87 87 88 88 90 90
87 80 88 88 88 88 91 93 92 92
89 89 89 88 79 79 79 79 78 77
...

These days I'm developing machine learning algorithm using scikit-learn python to predict critical situations in patient. since i got other 12 parameters(risk factors) that should be train along with above heart rate patterns. since it is not good practice to train these heart rate pattern and other 12 parameters altogether, i just need to find a way make these time series data set(Heart rate pattern) in to single value using any mathematical technique. it may be vector but i don't know how to do it!python scikit-learn based sample code will be very much useful.

Answer 1

Do the other variables vary over time? How many datapoints for the heartbeats do you have per patient? It's hard to answer without more information. The duration the heart beats were captured over need to be the same for an adequate comparison. However, here's some thoughts:

1. Compute some summary statistics, like mean, median, mode, min and max over the time series.
2. Compute moving averages over the time series, such as simple moving average, and exponential moving averages.
3. Compute the standard deviation of each datapoint

That will give you multiple numbers for the time series, but that is likely to be better for your model as it more completely describes the time series.