Predicting a Time Series Based on Previous Data

I have data with recorded roll, pitch, and yaw orientations over a certain amount of time and I have calculated the standard deviation and plotted a histogram of the dataset. Based off of this data I want to create a prediction model. Additionally, I want to be able to randomly generate a dataset of roll, pitch, and yaw over a certain time based off the probability of my current data. Any suggestions on how to do this?